Data Schema Explorer

Browse every delivered column of the Canaria data products, grouped by delivery module, from delivery catalog 8.3. Select a product below to explore its columns, types, and coverage.

Coverage is the share of rows where the field is populated, measured on 6 Oct 2026 for job postings, company, public-company and Google Maps data; each product shows its own measurement date. A dash means coverage has not been measured for that field on the current release.

For accuracy benchmarks on every enriched field, see our data quality benchmarks. For definitions of terms like SOC and semantic deduplication, see the glossary.

Job Postings Data: 176 fields, 1,065,010,950 rows (catalog 8.3, measured 6 Oct 2026)

Module	Field Name	Type	Description	Example	Coverage
jobs_core	jobID	string (fixed 64-char key)	SHA-256 per-source posting key: the primary key of the posting files and the join key between them. Changes if the employer edits the title; sourceWebsite with sourceJobKey is the title-edit-stable pair.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	100%
jobs_core	sourceJobKey	string, nullable	The source's own posting identifier, preserved exactly as the source published it. Stable when an employer edits the title; combine with sourceWebsite.	sWtlZMamILRhA845R7pY6Uc4FcPWv_gX9nB0Y-EiQdyOATcAQ-TCRQ	100%
jobs_core	sourceWebsite	string	The source the posting was collected from: aggregated job boards (for example indeed, linkedin, simplyhired, jora, postjobfree) and employer career portals (for example myworkdayjobs, greenhouse, lever, icims, smartrecruiters). Combine with sourceJobKey for joins that survive an employer editing the title. Some sources are retired: google (last posting 2024-08), careerbuilder and jobs2careers (2024-02), kariyer (2024-08) and jora (2025-10); their history stays. reference_source_coverage_monthly gives the monthly state (live, ramp, outage, retired) of every source with US postings.	simplyhired	100%
jobs_core	jobDate	datetime (UTC), nullable	Posting date for the role (UTC, millisecond precision): the source's published posting date when it supplies one, else the day we first collected the posting. jobDateSource says which; see Notes for using the data in the Guide before building daily or weekly series on this column.	2026-05-01 23:18:35.000	100%
jobs_core	jobDateSource	string	Where jobDate comes from: source (the source published it) or discovery_fallback (the source published none, so jobDate is the day we first collected the posting and equals firstSeenDate).	discovery_fallback	100%
jobs_core	firstSeenDate	datetime (UTC)	First time Canaria observed this posting (UTC, millisecond precision), except where firstSeenFromBackfill = 1.	2026-05-01 23:18:35.000	100%
jobs_core	firstSeenFromBackfill	integer	1 when firstSeenDate is not our observation but the posting's own date, taken from an archive to fill a 2024 gap in our PostJobFree collection. Such a row was not known to us on that date: leave it out of point-in-time work keyed on firstSeenDate.	0	100%
jobs_core	lastSeenDate	datetime (UTC)	Most recent time we observed this posting as still active (UTC, millisecond precision): the later of our last visit to the posting's page and the last time a search-results page or employer career portal still listed it. A listing never moves firstSeenDate. A duration computed from it is a lower bound on how long the posting was open.	2026-05-02 22:21:55.000	100%
jobs_core	contentID	string (fixed 64-char key), nullable	SHA-256 content-version hash over title + description. Shifts when either changes.	e296b88342730dfa05e7e7b5c29ca9c77e4d8187d3012639234abe9ad07e76b7	92.7%
jobs_core	isUniqueJobPosting	integer	1 if this posting represents its group of duplicates (the earliest seen), 0 if it duplicates another. SUM(isUniqueJobPosting) gives unique volume; COUNT() gives gross volume. A relisting first seen more than 60 days after the group's latest sighting starts a new unique posting; shorter gaps keep chaining. Postings that could not be tested count as unique. See Gross and unique postings in the Guide.	1	100%
jobs_core	duplicateGroupID	string (fixed 64-char key), nullable	Near-duplicate group key: postings sharing it are the same role posted on several sources or collected again over time, and the one with isUniqueJobPosting = 1 is the group's representative. NULL when the posting could not be tested for duplicates, because its location was not resolved precisely enough to compare it.	1de0b412ca7d479c32e343b2f80a6aafc6b27f6e2decff6ff8af342421a5aa98	>99.9%
jobs_core	duplicateGroupSize	integer, nullable	Number of postings in this near-duplicate group (repost intensity). 1 for a posting with no duplicates; NULL when de-duplication has not been applied.	3	>99.9%
jobs_core	representativeJobID	string (fixed 64-char key), nullable	The jobID of the group's representative posting (the earliest-seen instance, the one with isUniqueJobPosting=1). Points into the same key space as jobID, so a duplicate row joins straight back to its canonical row on representativeJobID = jobID. Equal to the row's own jobID for a representative; NULL when de-duplication has not been applied.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	>99.9%
jobs_core	isDedupProvisional	integer	Deduplication disclosure flag: 1 when the posting's near-duplicate block was left unresolved (block exceeded the resolution cap), so its duplicate grouping is provisional. Exclude flagged rows from deduplication-sensitive metrics for strict analyses.	0	100%
jobs_core	jobTitle	string, nullable	Raw posting title as published by the employer.	Health & Beauty TL	>99.9%
jobs_core	jobURL	string, nullable	Direct URL to the posting on the source site. The customer-clickable link to the job.	https://www.simplyhired.com/job/sWtlZMamILRhA845R7pY6Uc4FcPWv_gX9nB0Y-EiQdyOATcAQ-TCRQ	99.3%
jobs_core	atsApplicationURL	string, nullable	URL to the employer's ATS application form (Greenhouse, Lever, Workday, etc.) when the posting links out to one. May differ from jobURL when the posting lives on an aggregator. NULL when the only link the source gives is back to the job board itself (for example a LinkedIn posting's own linkedin.com/jobs/view/.. page): that is the board, not an application form.	https://careers.smartrecruiters.com/Dominos	2.9%
jobs_core	companyName	string, nullable	Company name as published on the posting, before normalization. companyName never has leading or trailing whitespace (including non-breaking spaces); a name that is empty after trimming, or HTML markup that cannot be stripped cleanly, is NULL.	Walmart	98.3%
jobs_core	companyProfileURL	string, nullable	URL of the employer's profile or company page on the source site (e.g. an Indeed company page, a LinkedIn company URL), when available.	https://www.linkedin.com/company/travel-nurse-source	27.0%
jobs_core	postedJobFunction	string, nullable	The job function the source site listed for the posting, exactly as stated (for example "Information Technology"). The source's own label, not a Canaria classification; it ships beside socCode, never in place of it. Filled mainly on LinkedIn postings and a few career portals; NULL where the source lists none.	Health Care Provider	18.7%
jobs_core	postedIndustry	string, nullable	The industry the source site listed for the posting's employer, exactly as stated (for example "Staffing and Recruiting"). The source's own label for this posting, so it can differ between postings of one company; it ships beside the NAICS columns, never in place of them. Filled mainly on LinkedIn and a few career portals; NULL where the source lists none.	Hospitals and Health Care	20.7%
jobs_core	postedDepartment	string, nullable	The department or team the source listed for the posting, as stated. A department list sent as a JSON array ships as the department names joined by " \| ", and a trailing " /" separator is removed; every other value is exactly as stated. Filled mainly on employer career portals; NULL where the source lists none.	Clinical	0.2%
jobs_core	postingExpirationDate	date, nullable	Reserved for a later release; empty on every row of this release.		0%
jobs_core	isUrgentHire	integer, nullable	Indeed postings only: 1 when Indeed marked the posting as urgently hiring, 0 when it did not. Indeed showed the mark on postings collected from 2022-04 to 2023-06 and from 2026-10; NULL elsewhere. New in catalog 8.3: it arrives with the next build.		-
jobs_core	isHighVolumeHiring	integer, nullable	Indeed postings only: 1 when Indeed marked the posting as high-volume hiring, 0 when it did not. Indeed showed the mark on postings collected from 2022-04 to 2023-06 and from 2026-10; NULL elsewhere. New in catalog 8.3: it arrives with the next build.		-
jobs_core	postingLanguage	string, nullable	ISO-style language tag detected on the description (en, es, fr, de, etc.).	en	92.1%
jobs_core	postingChannel	string	Where the posting was sourced: employer_ats (the employer's own applicant-tracking system, e.g. Workday/Greenhouse/Lever, so a direct employer), job_board (aggregator/marketplace such as Indeed/LinkedIn), or unknown.	job_board	100%
jobs_core	postedLocation	string, nullable	Location string exactly as the posting stated it, before parsing into the final columns. Source HTML markup is removed (each tag becomes a space, edges trimmed); NULL where removing the tags would leave a half-parsed string or nothing at all.	Moraine, OH	91.2%
jobs_core	finalCity	string, nullable	Resolved city name in normalized form.	Moraine	98.0%
jobs_core	finalState	string, nullable	Resolved state or top-level administrative region. US postings carry the 2-letter code (CA, WA); non-US postings carry the canonical region name.	OH	97.2%
jobs_core	finalCountry	string, nullable	Resolved ISO 3166-1 country, uppercase (US, GB, DE). US territories (Puerto Rico, Guam, US Virgin Islands, American Samoa, Northern Mariana Islands) are reported as US with the territory in finalState (PR, GU, VI, AS, MP).	US	99.2%
jobs_core	finalZipcode	string, nullable	Resolved postal code: the posting's own, when the source publishes one at posting level. A posting found by a location search without a ZIP of its own carries the search's ZIP only when that ZIP lies in the metro area (CBSA) of the city the posting names (or, for a city outside any metro, within 25 km of it); a posting that names no city we can place takes the search's ZIP as its only location.	45419	87.7%
jobs_core	finalLatitude	float, nullable	Latitude of the resolved location at zip/city precision. NULL when only a state- or country-level location could be resolved (we do not present a coarse centroid as a precise point).	39.70733642578125	97.8%
jobs_core	finalLongitude	float, nullable	Longitude of the resolved location at zip/city precision. NULL when only a state- or country-level location could be resolved.	-84.2156982421875	97.8%
jobs_core	cbsaCode	string, nullable	5-digit OMB Core-Based Statistical Area (metro/micro area) code for the posting's resolved ZIP, e.g. 35620. Assignment is one dominant CBSA per ZIP, derived by resolving the Census 2020 ZCTA-to-county crosswalk against the OMB July 2023 CBSA delineation. ZCTAs approximate the USPS ZIP inventory rather than reproducing it, so coverage is close to but not identical with the full ZIP list. A US posting with no resolved ZIP but a resolved city and state takes its city's CBSA: the one holding a strict majority of the city's ZIPs (a city split across metros gets none). NULL when the posting has neither a ZIP nor a mapped city, the ZIP lies outside any CBSA, the resolved country is non-US, OR the ZIP is absent from the ZCTA-derived source. A NULL is therefore "no metro mapping available", not a positive assertion that the location is rural.	19430	89.1%
jobs_core	cbsaTitle	string, nullable	Official OMB name of the CBSA in cbsaCode, e.g. New York-Newark-Jersey City, NY-NJ. NULL exactly when cbsaCode is NULL.	Dayton-Kettering-Beavercreek, OH	89.1%
jobs_role	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	100%
jobs_role	normalizedTitle	string, nullable	Canonical job title from Canaria's title taxonomy of about 100,000 roles: variant spellings, seniority prefixes and synonyms resolve to one canonical title.	Health and Beauty Clerk	99.8%
jobs_role	normalizedTitleConfidence	float, nullable	Confidence for normalizedTitle, 0 to 1, calibrated to P(correct).	0.28422087	99.8%
jobs_role	seniority	string, nullable	Seniority band classified from title and description: intern, entry, mid, supervisor, senior, staff_principal, director or executive. The classifier never declines: every posting it scores receives one of the eight bands, so a NULL is a posting no classifier scored.	entry	99.9%
jobs_role	seniorityTrack	string, nullable	Career track: ic (individual contributor), manager (people manager), or mixed. Distinguishes IC from manager at the same band.	ic	99.9%
jobs_role	seniorityConfidence	float, nullable	Confidence for seniority (the band only), 0 to 1. The classifier's own score, not calibrated to P(correct): choose a cut from this column's own distribution. socConfidence, by contrast, is calibrated and can be thresholded as a probability.	0.8919606	99.9%
jobs_role	seniorityDetail	string, nullable	Finance/banking sub-band refinement for senior roles (banking_vp, vp, c_suite, banking_md). NULL when not applicable. Disambiguates what "VP" means in banking vs tech.	vp	0.3%
jobs_role	seniorityDetailConfidence	float, nullable	Classifier score for seniorityDetail, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever seniorityDetail is NULL. seniorityTrack ships without a score of its own.	0.9698889	<0.1%
jobs_role	socCode	string, nullable	US SOC 2018 6-digit occupation code, classified from title and description context.	39-5012	99.9%
jobs_role	socTitle	string, nullable	Short label for the SOC code in socCode.	Hairdressers, Hairstylists, and Cosmetologists	99.9%
jobs_role	socConfidence	float, nullable	Confidence for socCode, isotonic-calibrated to P(correct), so it is meant to be read as a probability and thresholded as one. A declined classification is reported differently: socCode, socTitle and this column are NULL together, so a NULL means no occupation was assigned, never an assignment at zero confidence.	0.1808818	99.9%
jobs_arrangement	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	100%
jobs_arrangement	employmentStatus	string, nullable	Employment type: full_time, part_time, contract, temporary, internship, per_diem, volunteer, apprenticeship, residency_fellowship or other. A posting titled as a paid senior officer (vice president, president, a chief officer) never ships as volunteer: where the classifier says volunteer, the field is NULL. other is the classifier's rare catch-all (under 0.01% of postings, in practice executive roles with no stated arrangement).	contract	99.9%
jobs_arrangement	employmentConfidence	float, nullable	Classifier score for employmentStatus, 0 to 1, not calibrated to P(correct); the classifier never abstains, so choose a cut from this column's own distribution. NULL exactly when employmentStatus is NULL.	0.61192876	99.9%
jobs_arrangement	workerClassification	string, nullable	Staffing-market worker classification: w2, 1099, statutory_employee or c2c; in practice a w2 / 1099 split; statutory_employee (about 0.01% of postings, usually at low confidence) and c2c (a handful of postings) are emitted rarely and are not usable for analysis. NULL where the posting names no arrangement: the classifier answers unknown rather than default a contract posting to 1099. For postings that accept several arrangements, corp-to-corp included, see workerClassificationAccepted. Its own score is workerClassificationConfidence.	w2	99.1%
jobs_arrangement	workerClassificationConfidence	float, nullable	Classifier score for workerClassification, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever workerClassification is NULL.	0.9929842	1.1%
jobs_arrangement	workerClassificationAccepted	array<string>	For staffing postings that accept multiple arrangements ("W2/1099/C2C OK"), the multi-hot set of accepted classifications. Empty array when the posting names a single arrangement or gives no signal.	[]	0.2%
jobs_arrangement	industryEmploymentSubtype	string, nullable	Healthcare staffing subtype, classified from the posting text: travel, locum_tenens, telehealth_contractor, research_staff or staff. NULL when not a healthcare staffing posting. Because it reads the text, it also fires on some non-clinical travel roles (travelling merchandisers, field service technicians): combine it with a healthcare socCode (29- or 31-) for a clinical-only count.	travel	12.3%
jobs_arrangement	industryEmploymentSubtypeConfidence	float, nullable	Classifier score for industryEmploymentSubtype, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever industryEmploymentSubtype is NULL.	0.82015777	0.1%
jobs_arrangement	temporarySubtype	string, nullable	For temporary roles (employmentStatus = 'temporary'), the temporary nature (seasonal, project, leave_coverage, fixed_term, unspecified). NULL otherwise.	seasonal	1.4%
jobs_arrangement	temporarySubtypeConfidence	float, nullable	Classifier score for temporarySubtype, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever temporarySubtype is NULL.	0.8829124	<0.1%
jobs_arrangement	isGovernment	integer, nullable	1 if the employer is a government body, 0 / NULL otherwise. Pairs with requiresSecurityClearance.	0	99.9%
jobs_arrangement	isGovernmentConfidence	float, nullable	Classifier score for isGovernment, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever isGovernment is NULL.	0.920121	1.1%
jobs_arrangement	hasFractional	integer, nullable	1 if the role is fractional (e.g. fractional CFO / CMO), 0 / NULL otherwise. Deterministic text-pattern extraction (fractional / interim / retainer-based executive phrasing), not a model score.	0	99.9%
jobs_arrangement	hasDolApprentice	integer, nullable	1 if the posting is a US DOL-registered apprenticeship, 0 / NULL otherwise. Deterministic text-pattern extraction (DOL registration / apprenticeship.gov / 29 CFR phrasing), not a model score.	0	99.9%
jobs_arrangement	hasAcgmePgy	integer, nullable	1 if the posting is an ACGME residency / PGY-N training position, 0 / NULL otherwise. Deterministic text-pattern extraction (PGY-N / ACGME / residency-program phrasing), not a model score.	0	99.9%
jobs_arrangement	remoteStatus	string, nullable	Remote-work classification: Remote, Hybrid or On-Site. NULL only on postings no classifier scored.	On-Site	99.9%
jobs_arrangement	remoteConfidence	float, nullable	Classifier score for remoteStatus, 0 to 1, not calibrated to P(correct); the classifier never abstains. Values concentrate very high, so a 0.95 cut on this column filters far less than the round number suggests.	0.99875367	99.9%
jobs_arrangement	hybridDaysPerWeek	float, nullable	In-office days per week for hybrid roles, rounded to a whole day (0–7) so it filters cleanly (e.g. = 3). Populated only when remoteStatus = 'Hybrid'; NULL otherwise.	3	7.5%
jobs_arrangement	hybridSchedule	string, nullable	For hybrid roles, whether the in-office schedule is fixed, flexible or unspecified (the posting was read and states none). NULL on non-hybrid rows, and on the few hybrid rows not yet scored for a schedule.	fixed	7.5%
jobs_arrangement	hybridScheduleConfidence	float, nullable	Classifier score for hybridSchedule, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever hybridSchedule is NULL.	0.54035014	3.3%
jobs_arrangement	onsitePattern	string, nullable	For on-site roles, the work-location pattern: fixed_location, rotating (locum tenens / travel nursing), or mobile (field trades). NULL otherwise.	fixed_location	86.3%
jobs_arrangement	onsitePatternConfidence	float, nullable	Classifier score for onsitePattern, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever onsitePattern is NULL.	0.9484171	0.8%
jobs_arrangement	geoRestrictionLevel	string, nullable	Geographic hiring restriction granularity: none, country, state_or_region, metro, or timezone. NULL when there is no signal. Applies to any work mode.	none	99.9%
jobs_arrangement	geoRestrictionLevelConfidence	float, nullable	Classifier score for geoRestrictionLevel, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever geoRestrictionLevel is NULL.	0.90548044	4.3%
jobs_arrangement	hasFiveDayMandate	integer, nullable	1 if the posting states an explicit five-days-a-week in-office mandate, 0 or NULL otherwise. A text-pattern extraction (five-days-in-office phrasing), not a model score; a mandate also sets remoteStatus to On-Site, never Hybrid.	0	99.9%
jobs_arrangement	shiftTypes	array<string>	Shift-type phrases extracted (night shift, weekends, rotating shift, etc.).	[]	26.5%
jobs_arrangement	travelRequirements	array<string>	Travel-requirement phrases extracted (up to 25% travel, frequent client travel).	[]	4.2%
jobs_arrangement	shiftPattern	string, nullable	The posting's shift pattern as one value, mapped from shiftTypes: multiple (more than one fixed shift named), day, evening, night, rotating, split, shift_work (shifts, which not stated), compressed or flexible. NULL where the posting names none; 24/7 and weekend mentions alone are not mapped. Recall is limited: many postings that state a schedule in free prose get no value, so read a NULL as not detected rather than no shift.	flexible	15.6%
jobs_arrangement	travelRequirement	string, nullable	How much the job travels, as one value from travelRequirements and any travel percentage the description states: frequent, occasional, required (travel is part of the job, extent not stated) or none (the posting says no travel). NULL where the posting says nothing about travel.	required	4.6%
jobs_arrangement	startTiming	string, nullable	The canonical start-timing answer, fused from the three start-date extractors with priority parsed date, then signal, then label: specific_date (a literal date was parsed; see parsedStartDates), within_1_week, within_2_weeks, flexible, immediate (ASAP folds in). NULL when no start signal exists.	immediate	15.3%
jobs_arrangement	startTimingSource	string, nullable	Which extractor produced startTiming: parsed_date, signal, or label.	label	15.3%
jobs_arrangement	advertiserType	string	Type of advertiser (Direct Employer / Staffing Agency / staffing sub-types). Resolved strongest signal first: (1) Staffing Agency whenever a curated structured company signal identifies a staffing/recruiting firm (review-inferred inferredEntityType = staffing_agency, or its normalized industry is Staffing and Recruiting / Temporary Help Services, or, for an employer with no entity-type annotation (inferredEntityType = unknown), a staffing name: staffing, executive search, search partners, search group or headhunter in the name, or a recruiting word (recruiting, recruitment, recruiter, personnel, placement, talent, manpower, temps, temporaries, employment agency, employment services, interim, workforce) together with a model industry code of 561311 or 561320 at a naicsConfidence of 0.70 or better) AND the posting text is either silent or says Direct Employer; (2) otherwise the value inferred from posting text, if any; (3) Direct Employer when the text was silent and postingChannel = employer_ats; (4) Direct Employer when none of the above fired and the company-grain signal identifies a real operating employer (inferredEntityType = real_employer at an inferredEntityTypeConfidence of 0.70 or better, the floor below which the classification treats the label as unverified). Undetermined when no rung finds evidence either way: no advertiser phrase in the text, not an employer applicant-tracking-system feed, and an employer that is neither a known staffing company nor a confidently labelled direct employer. It is not defaulted to Direct Employer. Full domain: Direct Employer, Staffing Agency, Search/Recruiting Firm, Temp/Contract Staffing Agency, Recruitment Process Outsourcing, Employer of Record, Professional Employer Organization, Undetermined. Rung (1) is the only one that overrides a value a weaker rung produced, and it overrides exactly one value: a staffing-family text value (Staffing Agency, Temp/Contract Staffing Agency, Search/Recruiting Firm, Professional Employer Organization, Employer of Record, Recruitment Process Outsourcing) is never overridden, and no posting on a non-staffing company is affected. Consequence you can rely on: this column never reads Direct Employer on a posting whose inferredEntityType is staffing_agency. The company-grain rungs read the same signal in both directions, so the staffing share computed from this column tracks the one computed from inferredEntityType.	Direct Employer	100%
jobs_skills_extracted	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	100%
jobs_skills_extracted	technicalSkills	array<string>	Technical skills extracted from the description, as the skills taxonomy's canonical labels (Python (Programming Language), Node.js (Javascript Library), Terraform, React, etc.). An ambiguous name carries a qualifier in parentheses, so match the FULL label: has(technicalSkills, 'Python (Programming Language)'), never 'python'. Generic occupational nouns that are not skills (e.g. Operations, Scheduling, Transportation) are removed. A small set of domain terms (Finance, Accounting, Logistics, Transportation, Marketing, Construction) is retained only on relevant occupational roles, and dropped otherwise.	[]	86.5%
jobs_skills_extracted	softSkills	array<string>	Soft skills extracted, as canonical labels (Communication, Leadership, Attention to Detail).	[]	68.6%
jobs_skills_extracted	certifications	array<string>	Certifications named in the posting, as canonical labels: an abbreviation ships under the full name (CPA as Certified Public Accountant, PMP as Project Management Professional Certification), often with the abbreviation in parentheses (Registered Nurse (RN), Basic Life Support (BLS) Certification).	[]	25.8%
jobs_skills_extracted	licenseRequirements	array<string>	Professional license requirements extracted, as canonical labels (Driver's License, RN License, CDL Class A).	[]	21.2%
jobs_skills_extracted	degreeLevels	array<string>	Degree levels the description requires (High School, Associates, Bachelors, Masters, Doctoral, Certificate, Postbaccalaureate Certificate, Postmasters Certificate). A level is read only from degree wording, never from benefit text such as tuition assistance, and a level the occupation never requires is removed unless the text clearly states it as a credential. minimumDegree holds the lowest.	[]	38.7%
jobs_skills_extracted	minimumDegree	string	The lowest degree level required, collapsed from degreeLevels; both ship in this file. The collapse is lossy, so read degreeLevels when more than a floor matters. Empty string when no degree level is mentioned. Overloaded abbreviations are excluded where nothing disambiguates them: "dba" never emits Doctoral, and "j.d." emits it only in a degree shape ("J.D. required"), never in "J.D. Power".	High School	38.7%
jobs_skills_extracted	experienceLevels	array<string>	Experience-level cues mentioned (entry-level, senior, executive). Senior is dropped when the posting's seniority signal confidently indicates an entry or mid-level role, removing false positives such as "senior care" on non-senior postings.	[]	11.4%
jobs_skills_extracted	experienceRequirements	array<string>	Years-of-experience requirements parsed from the description, emitted as canonical labels (5+ years, 2-4 years, 10 years), not verbatim source spans: ranges normalize to N-M years and lead-ins like "minimum"/"at least" are stripped. Not converted to integers. Age minimums ("18 years of age") are excluded.	[]	18.5%
jobs_skills_extracted	languageRequirements	array<string>	Language-requirement phrases, including English when explicitly required (bilingual Spanish, fluent in Mandarin, fluent in English).	[]	6.1%
jobs_skills_extracted	physicalRequirements	array<string>	Physical-requirement phrases (lift 50 lbs, stand for 8 hours).	[]	16.6%
jobs_skills_extracted	employerTaggedSkills	array<string>	The skills the SOURCE listed for the posting as structured tags (Indeed's attribute tags, and the same field where another feed publishes it), exactly as listed: one label, or each element of the source's list. [] where the source listed none. Kept apart from technicalSkills / softSkills, which are extracted from the description and normalized to the taxonomy, so a buyer can tell what the employer tagged from what the text says. Not normalized: two sources can spell one skill differently.	[]	1.2%
jobs_skills_extracted	employerTaggedCertifications	array<string>	The certifications and licenses the source listed for the posting, as employerTaggedSkills describes. Compare certifications, extracted from the description.	[]	0.6%
jobs_skills_extracted	employerTaggedQualifications	array<string>	The qualifications the source listed for the posting (Indeed's qualification tags), as employerTaggedSkills describes.	[]	1.0%
jobs_salary	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	100%
jobs_salary	salaryString	string, nullable	The salary text the values were parsed from, as the posting showed it. NULL when the posting states no pay or the values are an estimate. A posting whose pay text gave no usable figure keeps the text beside NULL values: test the value columns, not the text, for whether a salary shipped.	$19 - $32 an hour	41.8%
jobs_salary	salarySource	string, nullable	Where the salary came from: description (stated in the posting text), source_field (the job board's structured salary field, as published; a board can fill that field with its own estimate without flagging it, so description is the only tier certainly in the employer's words), source_estimate (the job board's own flagged estimate) or predicted (a Canaria model estimate, where the posting states no pay). NULL when no salary value ships. Most resolved salaries are estimates, so read this column on every row.	scraped	90.5%
jobs_salary	salaryPayPeriod	string, nullable	The pay period the employer stated, which the annualSalaryUSD values were annualized from: hour, year, week, month, day, two_week or half_month, and rarely a per-unit rate such as shift or mile. Set only when salarySource is description or source_field; NULL on predicted and source_estimate rows (those are annual estimates, not a stated period), wherever salarySource is NULL, and on the rare stated salary whose text named no period. salaryPayPeriod = 'hour' selects every hourly-stated posting without parsing salaryString.	hour	38.5%
jobs_salary	annualSalaryMinimumUSD	float, nullable	Parsed annualized salary minimum, converted to USD at jobDate's FX rate. NULL if no lower bound parsed, the currency isn't covered, or the date is unknown.	39520	89.5%
jobs_salary	annualSalaryAverageUSD	float, nullable	USD-converted average.	53040	90.5%
jobs_salary	annualSalaryMaximumUSD	float, nullable	USD-converted maximum.	66560	88.7%
jobs_salary	salaryConfidence	float, nullable	For salarySource = 'predicted': the model's own confidence, 0 to 1, derived from its predictive uncertainty. It orders predictions but is not a success rate, and it is not the width of the salary range. On the other tiers it marks provenance: 1.0 for stated pay (description, source_field), 0.9 for job-board estimates (source_estimate). Select tiers with salarySource, not with this column (0.9 is stored as 0.8999999761581421, so an equality test matches nothing). NULL wherever no salary value ships.	1	90.5%
jobs_salary	salaryCurrencyCode	string, nullable	ISO 4217 currency code of the posting's original salary (USD, EUR, INR, KRW). Use this for filtering/grouping; the shipped annualSalaryUSD values are already converted to USD. NULL whenever no annualSalaryUSD value ships, including when the posting's currency had no exchange rate for its date.	USD	90.5%
jobs_salary	isPaid	integer, nullable	1 for paid roles, 0 for unpaid roles (unpaid internships, volunteer positions), NULL when there is no signal.	1	99.3%
jobs_salary	isPaidConfidence	float, nullable	Classifier score for isPaid, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever isPaid is NULL. Also NULL where a stated wage overrode the classifier: the score was for the class that did not ship.	0.9266869	1.1%
jobs_salary	benefits	array<string>	Employer-provided benefits (401K, PAID TIME OFF, HEALTH INSURANCE, etc.): those mentioned in the description, followed by any the job board listed for the posting that the description does not mention, under the same labels. A posting with no description carries the labels its job board listed. A board label that is not a benefit in this vocabulary ("Flexible schedule", "Remote", "Employee stock purchase plan") adds nothing here.	[]	54.2%
jobs_salary	equity	array<string>	Equity or stock-based compensation mentioned (ESPP, Equity, Stock Options, and the rare Cash or Share Options). Distinct from cash benefits in benefits.	[]	3.8%
jobs_salary	compensation	array<string>	Direct cash compensation beyond base pay mentioned (Commission, Performance Bonus, Overtime Pay, Shift Differential, Hazard Pay, Signing Bonus, Severance Pay). Wages, distinct from employer-provided benefits in benefits.	[]	8.7%
jobs_company	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	100%
jobs_company	normalizedCompanyName	string, nullable	Canonical employer name from Canaria's company database: the join key to the company and ticker files. Normally resolved from the posting's own companyName; where a posting names no employer, it can be recovered from its companyProfileURL, or carried from its duplicate group when every named member names the same company.	Walmart	98.2%
jobs_company	normalizedIndustryName	string, nullable	Canonical industry of the employer, one spelling across the variants sources use.	Retail	82.3%
jobs_company	naicsSector	string, nullable	NAICS 2022 sector (2-digit; the multi-prefix sectors read 31-33, 44-45 and 48-49), from whichever source naicsSource names.	44-45	98.0%
jobs_company	naicsIndustryGroup	string, nullable	NAICS-2022 industry group (4-digit). The first four digits of naicsCode wherever a 6-digit code resolved. Label it from the reference_naics bundle at naicsLevel = 4.	4552	98.0%
jobs_company	naicsCode	string, nullable	NAICS 2022 6-digit industry code. NULL mostly where the classifier declined to place the employer; a few of those postings still carry a sector and title from the curated industry lookup.	455211	98.0%
jobs_company	naicsTitle	string, nullable	Official NAICS 2022 label of the most specific level that resolved.	Warehouse Clubs and Supercenters	98.0%
jobs_company	naicsSource	string, nullable	Where both sources have a code, the model wins at or above 0.30 confidence, whether or not the two agree on the sector, and the curated lookup wins otherwise. NULL when no NAICS could be resolved.	industryMap	98.0%
jobs_company	naicsConfidence	float, nullable	Model confidence for the NAICS code, 0 to 1, calibrated to P(correct). Set on model rows; NULL on curated-lookup rows (the lookup carries no score) and where no NAICS resolved. Filter on naicsSource first.	0.9879	51.7%
jobs_company	sourceEntityType	string, nullable	Source-reported legal or organizational structure of the employer (Public Company, Privately Held, Nonprofit, Government Agency, Educational, Partnership, Self-Owned, Self-Employed). The source-reported companion to the classifier-inferred inferredEntityType.	Public Company	71.0%
jobs_company	companyRevenueMinimumUSD	float, nullable	Lower bound (inclusive) of the employer's annual revenue range, in USD.	10000000000	38.6%
jobs_company	companyRevenueMaximumUSD	float, nullable	Upper bound (inclusive) of the employer's annual revenue range, in USD; NULL for open-ended top buckets.	5000000000	28.2%
jobs_company	companyHeadquartersCountryISO2	string, nullable	ISO 3166-1 alpha-2 country code (uppercase) for the employer's headquarters country.	US	78.7%
jobs_company	companyHeadquartersCity	string, nullable	Employer headquarters city in normalized form.	Bentonville	78.3%
jobs_company	companyHeadquartersState	string, nullable	Employer headquarters state or top-level administrative region. US and Canadian employers carry the 2-letter code; others carry the canonical region name.	AR	73.0%
jobs_company	tickerSymbol	string, nullable	Ticker of the listed company the employer resolves to (e.g. AAPL): its own listing or its owning parent's. The link is decided once per employer name and listed company, from a reviewed sample of that name's postings, and applies to its postings inside the company's ticker window. NULL on a posting a marketplace rule identifies as a third party's listing on the company's platform, and on one whose link did not yet hold on its first-seen date (see the Guide on public-company links).	WMT	23.3%
jobs_company	isPubliclyTraded	integer, nullable	1 when the employer itself is the listed company (including under a former name), 0 when the employer is owned by a listed parent (subsidiary, acquisition, or division), or carries the linked issuer's ticker without being owned by it, NULL when no public-company link was resolved.	1	23.3%
jobs_company	stockExchange	string, nullable	Exchange the linked ticker trades on, one spelling per venue: NYSE, NASDAQ, NYSE AMERICAN, OTHER (NYSE Arca, Cboe BZX and IEX) or OTC (over-the-counter quotations) for US listings. NULL when tickerSymbol is NULL or the link records no venue. A symbol is not unique across venues: join on tickerSymbol and stockExchange.	NASDAQ	21.4%
jobs_company	tickerRelation	string, nullable	How the employer relates to the linked ticker: self (the listed company), former_name (the listed company under an earlier name), subsidiary, division or acquisition (owned by the listed company), franchise (an independent franchisee operating under the listed company's brand, no equity link), partner (a contracted operator under the brand, such as a delivery or freight partner, no equity link), fund_portfolio (a company held by funds a listed manager runs) or bdc_holding (a portfolio investment of a listed business development company).	self	23.3%
jobs_company	tickerParentName	string, nullable	Registered name of the company that owns the linked ticker (the employer itself for self and former_name links). NULL when tickerSymbol is NULL, and where the owning parent could not be named. Use tickerRelation and isPubliclyTraded, not this column, to tell a parent attribution from an own listing.	Walmart	21.5%
jobs_company	tickerMatchConfidence	float, nullable	Confidence score (0 to 1) of the employer-to-ticker link. Links ship at 0.85 and above. NULL when tickerSymbol is NULL.	1	23.3%
jobs_company	tickerMatchMethod	string, nullable	How the link was established: curated (a reviewed link), website (the employer's own domain matched the issuer's), name_exact, registrant_name (the issuer's registered filer name), name_variant, name_fuzzy, exact_alias (a reviewed alias that links only the one name it lists), corporate_family (inherited from a corporate parent or sibling link), curated_franchise (a reviewed franchise link) or brand_operator (a franchisee or partner whose own name pairs a listed brand with an operator, such as "DWMJ Inc d/b/a McDonald's"). Not every method occurs in every release; the highest-precision subset is curated, website and name_exact.	annotation	23.3%
jobs_company	isUSListing	integer, nullable	1 when the linked ticker trades on a US exchange (NYSE, NASDAQ, NYSE AMERICAN or OTHER (the bucket for NYSE Arca, Cboe BZX and IEX)), 0 when it trades outside the US, NULL when there is no ticker link, the link records no venue, or the venue is the over-the-counter market. Filter isUSListing = 1 for a US equity panel.	1	21.1%
jobs_company	primaryListingSymbol	string, nullable	The company's home / ordinary-share listing symbol, carried beside tickerSymbol rather than instead of it, and resolved as of this posting's own date like the rest of the ticker block. primaryListingSymbol != tickerSymbol means tickerSymbol is not the home listing: an ADR or another secondary line: a sufficient test, not an exhaustive one, and it must not be read as "everything else is a single listing". The home leg exists only where Canaria's curated listing reference happens to hold both legs of one company, and there is no other source for it: the SEC-derived listing-date reference holds neither leg of these pairs. So use the inequality to confirm a secondary line, never to rule one out. NULL when tickerSymbol is NULL.	WMT	23.3%
jobs_company	primaryListingExchange	string, nullable	Exchange primaryListingSymbol trades on, canonicalized to one spelling per venue exactly like stockExchange. NULL when tickerSymbol is NULL, and ALSO NULL whenever the link itself carries no venue: the two halves of the pair do not fill together.	NASDAQ	21.4%
jobs_company	publicFloatUSD	integer, nullable	The public float (market value of the common equity held by non-affiliates, in USD) of the SEC registrant behind tickerSymbol, from the latest annual report (10-K) filed on or before the posting's date: no report filed after the posting is ever used. NULL where tickerSymbol is NULL or no float applies. A subsidiary posting under its parent's ticker carries the parent's float. Floats tagged in the wrong unit are screened out; a small residue remains (see Notes for using the data in the Guide).	391703475732	18.9%
jobs_company	publicFloatAsOf	date, nullable	The date publicFloatUSD is measured at, as the report states it (the last business day of the issuer's second fiscal quarter). NULL exactly when publicFloatUSD is NULL.	2025-07-31	18.9%
jobs_company	marketCapTier	string, nullable	Market-cap tier from publicFloatUSD: large (10 billion USD or more), mid (2 to 10 billion), small (300 million to 2 billion) or micro (under 300 million). Point in time like publicFloatUSD, so usable as a backtest universe filter. The company files carry no tier, because a current tier on a historical posting would be lookahead.	large	18.9%
jobs_company	companyStatus	string	Ownership or sector class, always populated: public (the listed company itself), subsidiary_of_public (a business it owns), franchisee_of_public or partner_of_public (a franchisee or contracted operator under its brand, no equity link), portfolio_company_of_public (held by funds a listed manager runs), government, nonprofit, educational, private, or unknown. A verified listing outranks the reported sector. It always agrees with the ticker block on the same row: where the block is NULL, it reads the employer's class without a link.	public	100%
jobs_company	inferredEntityType	string	Hiring role of the company: real_employer, staffing_agency, job_board_aggregator (a job board, not an employer), not_a_company, or unknown. A role ships only at 0.70 confidence or more, so unknown means not classified, not doubtful. Sources, in order of precedence: a reviewed classification, a verified stock-ticker link, the source-reported organisation type (Nonprofit, Educational), and a classifier reading the company's own postings. A separate axis from sourceEntityType (legal form) and companyStatus (ownership).	real_employer	100%
jobs_company	inferredEntityTypeConfidence	float, nullable	Per-field confidence (0-1) of inferredEntityType. NULL exactly when inferredEntityType is unknown. For a classified employer this is the score recorded by the model-based company classification; for an employer resolved by the ticker-link rule it is that rule's measured precision (0.95); for one resolved by the organisation-type rule it is that label's measured precision (0.987 for Nonprofit, 0.955 for Educational); for one resolved by the posting-evidence classifier it is that classifier's own score, 0.70 or more.	1	97.0%
jobs_company	attributesAsOf	date	The UTC day this posting's company attributes (industry name, NAICS, size and revenue bands, headquarters, entity type) were resolved. They are that day's values for the employer, not values as of the posting's date: profiles exist only from 2024-09 on, so a 2022 posting carries its employer as resolved now. The ticker block, public float and market-cap tier are point in time and are not covered by this date.	2026-10-06	100%
jobs_signals	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	100%
jobs_signals	remoteWorkMentions	array<string>	Remote-work phrases extracted (fully remote, work from home, hybrid 3 days in office, etc.).	[]	12.6%
jobs_signals	parsedRemotePercentages	array<string>	Remote-percent strings parsed from the description (verbatim spans, e.g. 3 days in office, 100% remote).	[]	0.7%
jobs_signals	remotePercentageLabels	array<string>	Qualitative remote tags (fully remote, fully onsite, remote-first). Independent of parsedRemotePercentages, not a per-entry pairing.	[]	0.8%
jobs_signals	parsedStartDates	array<string>	Start-date strings parsed from the description (verbatim spans).	[]	4.4%
jobs_signals	startDateLabels	array<string>	Qualitative start-date tags (immediate, ASAP, flexible). Independent of parsedStartDates, not a per-entry pairing.	[]	11.5%
jobs_signals	startDateSignals	array<string>	Start-timing phrases from the signal extractor (Flexible Start, Start Within 1 Week, Start Within 2 Weeks). Finer-grained vocabulary than startDateLabels; both feed the canonical startTiming fusion.	['Start Within 2 Weeks']	0.1%
jobs_signals	employmentTypeMentions	array<string>	Employment types mentioned in the text (full-time, part-time, contract, internship, etc.). Verbatim mentions; employmentStatus is the single classified value.	[]	59.3%
jobs_signals	urgencySignals	array<string>	Urgency-signal phrases (immediate start, urgent hiring, start ASAP).	[]	7.8%
jobs_signals	companyTypes	array<string>	Company-type cues mentioned in the posting (startup, fortune 500, non-profit).	[]	16.5%
jobs_signals	eeocStatement	array<string>	EEOC / equal-opportunity statement categories detected on the posting (one element per statement; a posting can carry several). Empty array when none.	[]	37.6%
jobs_eligibility	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	100%
jobs_eligibility	offersVisaSponsorship	integer, nullable	1 if the posting says the role offers visa sponsorship, 0 if it says it does not, NULL when it says neither. Read from the posting text: application questions and bare headings are not statements, and where a posting both refuses and appears to offer, the refusal wins.	0	1.9%
jobs_eligibility	requiresSecurityClearance	integer, nullable	1 if the posting states that the role requires a security clearance, 0 if it states that it does not, NULL when the posting makes no such statement. Read from the posting text only, so a NULL is "the posting does not say", never "no clearance needed".	1	1.5%
jobs_eligibility	citizenshipRequirements	array<string>	Citizenship-requirement phrases extracted from the posting.	[]	5.8%
jobs_eligibility	screeningRequirements	array<string>	Screening-requirement phrases (background check, drug test, credit check).	[]	10.3%
jobs_eligibility	visaSponsorshipMentions	array<string>	Visa-sponsorship phrases extracted from the posting (mentions of H-1B, sponsorship language, etc.).	[]	1.5%
jobs_eligibility	clearanceLevels	array<string>	Security clearance levels named in the posting, as canonical labels; match on the full label. Keyword-extracted and independent of requiresSecurityClearance, so a level can appear on a posting that states no requirement.	[]	1.5%
jobs_contacts	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	100%
jobs_contacts	extractedEmails	array<string>	Email addresses found in the description text.	['careers@example-vet.com']	33.4%
jobs_contacts	extractedPhones	array<string>	Phone numbers found in the description text.	['+1-555-0142']	41.4%
jobs_contacts	extractedURLs	array<string>	URLs found in the description text (distinct from jobURL, the posting's own link).	[]	49.7%
jobs_descriptions	contentID	string (fixed 64-char key), nullable	SHA-256 content-version hash over title + description. Shifts when either changes.	3cffff7ad7d58aaa3356e53ee724dd9dfbc3999657146d03b9561a611e80356f	100%
jobs_descriptions	jobDescriptionCleaned	string	Cleaned full job description text: boilerplate and markup stripped, whitespace normalized. Heavy text is split into this file so jobs_core stays narrow; join to jobs_core on contentID.	Our growing veterinary practice is seeking an associate veterinarian to join our team...	100%
jobs_clinical	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.	f300000caf40825596634d529102a76d61157f94c6ae2fbe2d2b5cc55c7bd153	100%
jobs_clinical	complianceRequirements	array<string>	Clinical / regulatory compliance phrases extracted from the description (e.g. HIPAA, CMS, Joint Commission).	[]	11.1%
jobs_clinical	facilityTypes	array<string>	Facility-type mentions extracted from the description (hospital, outpatient clinic, long-term care, skilled nursing facility).	['Community Hospital', 'Magnet Hospital', 'Level II Trauma Center']	55.3%
jobs_clinical	patientPopulations	array<string>	Patient-population mentions (pediatric, geriatric, oncology, behavioral health).	['Pediatric']	52.0%
jobs_clinical	caseloads	array<string>	Caseload-related phrases extracted from the description (patient volumes, ratios).	[]	11.0%
jobs_clinical	credentialingTimelines	array<string>	Credentialing-timeline phrases extracted from the description (onboarding, license verification windows, start-readiness).	[]	0.4%
jobs_changes	jobID	string (fixed 64-char key)	The posting's key; joins to jobs_core.		-
jobs_changes	changeType	string	'upsert' (the posting is new or its delivered values changed; its full current row is in the other jobs files) or 'removal' (delete it on jobID). A full delivery lists every live posting with its latest change and no removals.	upsert	-
jobs_changes	changedAt	datetime (UTC)	When the change was recorded (UTC).		-
jobs_changes	vintage	string	The release the change came from.		-
jobs_open_duration	jobID	string	The posting's key; joins to jobs_core.	deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921	-
jobs_open_duration	postingStatus	string	closed, open (not yet unlisted for its threshold: still open, or closed too recently to tell), or unmeasurable.		-
jobs_open_duration	unmeasurableReason	string, nullable	Set only when unmeasurable: not_list_crawled (no listing page ever showed the posting: sources collected from detail pages only, most of 2022-2023, and Indeed from 2025-04 to 2026-07), crawl_stopped (we stopped collecting its source or board before it could be confirmed gone), insufficient_crawl_history (no collection calendar or threshold for its source in this release) or not_yet_measured (published after this release's measurement).		-
jobs_open_duration	lastListedDate	date, nullable	The last day a listing collection showed the posting.		-
jobs_open_duration	postingOpenDays	integer, nullable	Calendar days from firstSeenDate to lastListedDate, both included. For closed this is the open duration; for open, the days open so far (right-censored). NULL when unmeasurable.		-
jobs_open_duration	postingDaysSeen	integer, nullable	Distinct days any of our collection saw the posting.		-
jobs_open_duration	sightingCoverage	float, nullable	postingDaysSeen over the collected days of its unit between first and last sighting, capped at 1. Below 0.5 the span rests on few sightings: treat postingOpenDays as low confidence. NULL where it cannot be computed.		-
jobs_open_duration	unlistedCrawledDays	integer, nullable	Collected days of its unit since lastListedDate, through measuredThroughDate. A stricter reader can require more than closureThresholdDays.		-
jobs_open_duration	closureThresholdDays	integer, nullable	Collected days a posting must go unlisted before it counts as closed. Set per source from that source's own history (the smallest gap after which fewer than 5% of postings reappear, between 3 and 60 days) and re-measured on every release.		-
jobs_open_duration	closureConfidence	float, nullable	1 minus the share of the source's postings that were listed again after being unlisted this many collected days; 0 for a posting listed on its unit's last collected day. Set for open rows too.		-
jobs_open_duration	closureKnowableDate	date, nullable	For closed only: the first date on which the verdict could have been known: the day after the threshold's last collected day for an employer board; for a job board, the day after the 14th day of the source's calendar that follows it, because whether a job-board day counts as collected compares it with the 14 days either side.		-
jobs_open_duration	isLeftCensored	integer, nullable	1 when the posting may have opened before we first saw it, so postingOpenDays is a lower bound.		-
jobs_open_duration	leftCensorReason	string, nullable	posted_before_first_seen (a posting date from our historical archive precedes our first observation), before_list_crawling (first seen before its unit had a listing collection), list_crawling_start (first seen in its unit's first 7 collected days, when a newly collected board is already full of older postings).		-
jobs_open_duration	sourceLastCrawledDate	date, nullable	The last day its unit was collected.		-
jobs_open_duration	measuredThroughDate	date, nullable	The last collected day this release measured through: every verdict is as of this date.		-

Company Data: 143 fields, 29,114,505 rows (catalog 8.3, measured 6 Oct 2026)

Module	Field Name	Type	Description	Example	Coverage
company_profiles	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.	amazon	100%
company_profiles	affiliatedCompanyNames	array<string>	Distinct names of companies affiliated with this one on the source profile (e.g. LinkedIn affiliated pages). Empty array when none.	['AWS','Twitch']	0.4%
company_profiles	attributesAsOf	date	The UTC day this row's company attributes (industry, NAICS, size and revenue bands, headquarters and the rest) were resolved: the release day. They are the release's current values, not values as of any posting date.	2026-09-30	-
company_profiles	companyCEO	string	Chief executive's name where the source profile lists one.	Jane Doe	0.1%
company_profiles	companyDataSources	array<string>	Sources that contributed to this profile (array form of firmographicSources). Each firmographic field is chosen separately, preferring LinkedIn where both sources have a value, else the most recent observation, so one company's industry and size band can come from different sources.	['indeed','linkedin']	87.3%
company_profiles	companyDescription	string	Short company description.	More than 60 years ago, Sam Walton opened the first Walmart in Rogers, Arkansas. Today,...	73.5%
company_profiles	companyDomain	string	Registrable domain of companyWebsite, the buyer join key to web-domain datasets.	amazon.com	60.0%
company_profiles	companyEmployeeCount	integer	Employee count shown on the source company profile, where it shows one. It is sourced separately from the self-reported size band (companyEmployeeSizeBucket) and usually lower, because not every employee is listed: use the size band to size a company and this column for the observed count.	1540000	67.3%
company_profiles	companyEmployeeSizeBucket	string	Self-reported employee size band: 1, 2-10, 11-50, 51-200, 201-500, 501-1000, 1001-5000, 5001-10000 or 10001+.	1001-5000	73.5%
company_profiles	companyEmployeeSizeMaximum	integer	Upper bound (inclusive) of the size bucket; NULL for open-ended buckets like 10001+.	5000	73.3%
company_profiles	companyEmployeeSizeMinimum	integer	Lower bound (inclusive) of the size bucket as an integer. Useful for numeric range filters.	1001	73.5%
company_profiles	companyFollowerCount	integer	Follower count on the source company profile (e.g. LinkedIn). NULL when not disclosed.	2100000	64.7%
company_profiles	companyFoundedYear	integer	Year the company was founded.	1994	29.8%
company_profiles	companyHeadquartersCity	string	Headquarters city (canonical, normalized form).	Mountain View	67.8%
company_profiles	companyHeadquartersCountryISO2	string	ISO 3166-1 alpha-2 code (uppercase) of the headquarters country. A few profiles carry a retired or user-assigned code as the source gives it (AN, CS, XK).	US	73.4%
company_profiles	companyHeadquartersState	string	Headquarters state or top-level administrative region. US and Canadian companies carry the 2-letter code (CA, WA, ON); others carry the canonical region name. Resolved by the same location rules as a posting's finalState.	CA	45.3%
company_profiles	companyHeadquartersZipcode	string	Headquarters postal / ZIP code, parsed from the HQ location string. NULL when the source did not carry one.	98109	0.6%
company_profiles	companyLastVerifiedDate	timestamp	Most recent time a source profile behind this row was collected.	2026-09-22 12:00:00	13.3%
company_profiles	companyLinkedInURL	string	The company's LinkedIn page (where several match, the one with the most followers).		10.7%
company_profiles	companyLogoURL	string	URL of the company logo image from the source profile. NULL when not present.	https://…/logo.png	77.4%
company_profiles	companyOfficeLocations	array<string>	"City, State, Country" for each office listed on the company's source profiles; empty when none is listed.	[]	5.8%
company_profiles	companyRating	float	Aggregate company rating on the source profile (source scale). NULL when not rated.	4.1	2.6%
company_profiles	companyRevenueMaximumUSD	float	Upper bound (inclusive) of the revenue bucket in USD; NULL for open-ended buckets like more than $10B (USD).	500000000.0	0.4%
company_profiles	companyRevenueMinimumUSD	float	Lower bound (inclusive) of the revenue bucket in USD.	100000000.0	0.5%
company_profiles	companyRevenueRange	string	Annual revenue band in USD: less than $1M, $1M to $5M, $5M to $25M, $25M to $100M, $100M to $500M, $500M to $1B, $1B to $5B, $5B to $10B or more than $10B (each label ends in "(USD)"). Filled where a source profile states revenue.	$1B to $5B (USD)	0.5%
company_profiles	companyReviewCount	integer	Number of reviews behind companyRating. NULL when not disclosed.	18452	2.6%
company_profiles	companySpecialties	array<string>	Distinct self-declared specialties / focus areas from the source profile. Empty array when none.	['Cloud Computing','E-commerce']	3.4%
company_profiles	companyStatus	string	Ownership and sector class, set on every profile: public (the listed issuer), subsidiary_of_public (owned by a listed parent), franchisee_of_public, partner_of_public or portfolio_company_of_public (operating under, contracted to, or held by funds of a listed company that does not own it), government, nonprofit, educational, private, or unknown. A verified listing outranks the reported sector; a company identified as listed or listed-owned whose ticker could not be linked reads public or subsidiary_of_public with no tickerSymbol.	private	100%
company_profiles	companyWebsite	string	Primary company domain.	amazon.com	60.0%
company_profiles	firmographicSourceCount	integer	Number of source profile records that contributed to this row; more records mean better-corroborated firmographics.	6	100%
company_profiles	firmographicSources	string	Comma-separated source systems that contributed firmographics.	linkedin,indeed	87.3%
company_profiles	firstPostingDate	timestamp	Earliest jobDate of any posting attributed to this company (UTC). Where a source publishes no posting date, jobDate is the day we first collected the posting, so an early firstPostingDate can reflect when collection of that source began. A small number of companies carry a first posting before January 2022, from postings collected before the panel starts.	2022-01-14 00:00:00.000	29.9%
company_profiles	placesAvgRating	float	Average star rating (0-5) across matched Google Maps locations, weighted by each location's review count. NULL under the same rule as placesLocationCount, or when none of the matched locations carry a rating.	4.2	13.6%
company_profiles	placesLocationCount	integer	Number of Google Maps business locations matched to this company. Locations flagged third-party-sited are excluded from every places roll-up. NULL when no location matched or when every matched location was flagged.	1842	17.8%
company_profiles	placesNaicsCode	string	Most common NAICS-2022 industry code across the company's matched Google Maps locations (the modal value). NULL under the same rule as placesLocationCount.	722511	17.0%
company_profiles	placesReviewCount	integer	Sum of customer review counts across all matched Google Maps locations. NULL under the same rule as placesLocationCount.	983204	17.8%
company_profiles	placesTopCategories	array<string>	Most common Google Maps business categories across the company's matched locations, most frequent first. Empty when no location matched or every matched location was flagged.	['Coffee shop','Cafe']	17.7%
company_profiles	placesTotalWeeklyHours	float	Sum of weekly open hours across all matched Google Maps locations, a coarse proxy for physical-footprint scale. NULL under the same rule as placesLocationCount.	128640.0	13.3%
company_profiles	inferredEntityType	string	Hiring role of the company: real_employer, staffing_agency, job_board_aggregator (a job board, not an employer), not_a_company, or unknown. A role ships only at 0.70 confidence or more, so unknown means not classified, not doubtful. Sources, in order of precedence: a reviewed classification, a verified stock-ticker link, the source-reported organisation type (Nonprofit, Educational), and a classifier reading the company's own postings. A separate axis from sourceEntityType (legal form) and companyStatus (ownership).	real_employer	100%
company_profiles	inferredEntityTypeConfidence	float	Confidence (0-1) of inferredEntityType; NULL exactly when inferredEntityType is unknown. For a classified company it is the classifier's score, 0.70 or more; for a company resolved by the ticker-link rule, that rule's measured precision (0.95); for one resolved by the organisation-type rule, that label's measured precision (0.987 for Nonprofit, 0.955 for Educational).	0.94	20.0%
company_profiles	isPubliclyTraded	integer	1 when this company is the listed entity (including listings under a former name); 0 when it is owned by a listed parent (subsidiary, acquisition, division) or operates under one's brand; NULL when there is no link to a public listing.	1	0.2%
company_profiles	isUSListing	integer	1 when the linked ticker trades on a US exchange (NYSE, NASDAQ, NYSE AMERICAN or OTHER (the bucket for NYSE Arca, Cboe BZX and IEX)), 0 when it trades outside the US (for example XETRA, LSE, Euronext Paris, TSX). NULL when there is no ticker link, the link records no exchange, or the ticker trades over the counter (OTC). Filter isUSListing = 1 for a US equity panel.	1	0.1%
company_profiles	lastPostingDate	timestamp	Latest jobDate. Millisecond precision in UTC.	2026-04-01 00:00:00.000	29.9%
company_profiles	lifetimeDistinctSOCCount	integer	Distinct US Standard Occupational Classification codes across the company's lifetime.	42	100%
company_profiles	lifetimeDistinctStatesCount	integer	Distinct US states the company has hired in.	50	100%
company_profiles	lifetimePostingsCount	integer	Gross count of all job postings attributed to this company (every posting, before near-duplicate collapse). Approximate, within about 0.5%.	247830	100%
company_profiles	lifetimePostingsWithClearanceCount	integer	Lifetime postings whose description mentions a security clearance requirement.	380	100%
company_profiles	lifetimePostingsWithRemoteCount	integer	Lifetime postings flagged as remote.	38940	100%
company_profiles	lifetimePostingsWithSalaryCount	integer	Lifetime postings whose text or the source site's structured salary field states pay (the same rule as postingsWithAnyPostedSalaryCount in Hiring History). Structured fields can include a site's own estimate; for pay stated in the posting text, use Hiring History's postingsWithSalaryCount. Approximate, within about 0.5%.	142109	100%
company_profiles	lifetimePostingsWithVisaSponsorshipCount	integer	Lifetime postings whose description mentions visa sponsorship, whether offering or declining it.	5021	100%
company_profiles	lifetimeUniquePostingsCount	integer	De-duplicated lifetime posting count: unique postings only (isUniqueJobPosting = 1), so the same role posted on several sources or collected again over time counts once. Approximate, within about 0.5%.	198042	100%
company_profiles	nAffiliatedPages	integer	Count of distinct affiliated companies (see affiliatedCompanyNames).	12	100%
company_profiles	nCompanyLocations	integer	Count of distinct office locations listed on the source profile.	27	100%
company_profiles	nCompanyUpdates	integer	Count of distinct company posts / updates captured from the source profile.	340	100%
company_profiles	nEmployeesListed	integer	Count of distinct employees surfaced across the company's source profiles.	1240	100%
company_profiles	nSimilarPages	integer	Count of distinct similar companies (see similarCompanyNames).	20	100%
company_profiles	naicsCode	string	NAICS 2022 6-digit industry code. NULL where a company can be placed in a sector but not a 6-digit industry (for example "Manufacturing", sector 31-33).	513210	100%
company_profiles	naicsConfidence	float	Model confidence in naicsCode (0-1), calibrated to the probability the code is correct; set only when naicsSource is model. For companies with few records the score runs low, so the code is right more often than it says: set thresholds by company size rather than one cut for the whole file.	0.71	89.0%
company_profiles	naicsIndustryGroup	string	NAICS-2022 industry group (4-digit), the first four digits of the winning code. Populated when the winning source is specific enough to land at the 4-digit level.	5132	100%
company_profiles	naicsSector	string	NAICS-2022 sector code (2-digit, with canonical range form 31-33/44-45/48-49 for multi-prefix sectors). Comes from whichever NAICS source won for this company: a reviewed public-issuer or large-employer code, the model or the curated lookup. NULL when no source places the company.	51	100%
company_profiles	naicsSource	string	Which side of the resolution above produced naicsCode / naicsSector / naicsIndustryGroup / naicsTitle: model for the classification model, industry_lookup for the curated lookup. NULL when no NAICS was resolved at all.	model	100%
company_profiles	naicsTitle	string	Short label for the most-specific populated NAICS level.	Software Publishers	100%
company_profiles	normalizedIndustryName	string	Canonical industry label, one spelling across the variants sources use. Filter or join on this.	Information Technology & Services	73.5%
company_profiles	primaryListingExchange	string	Exchange primaryListingSymbol trades on, canonicalized to one spelling per venue exactly like stockExchange. NULL when tickerSymbol is NULL, and ALSO NULL whenever the link itself carries no venue: the two halves of the pair do not fill together, so a non-NULL primaryListingSymbol does not imply a venue beside it. Same contract as the jobs source's column of this name.	SIX SWISS	0.1%
company_profiles	primaryListingSymbol	string	The company's home / ordinary-share listing symbol, carried beside tickerSymbol rather than instead of it. primaryListingSymbol != tickerSymbol means tickerSymbol is not the home listing, an ADR or another secondary line, a sufficient test, not an exhaustive one (see the dual-listing note below for the measured coverage). Note isUSListing describes tickerSymbol, not this pair. NULL when tickerSymbol is NULL, matching the jobs source's column of this name.	ZURN	0.2%
company_profiles	relatedCompaniesCount	integer	Number of companies that share substantial job-posting content with this one, published with topRelatedCompanies. 0 when none are known; this release publishes no related companies, so it reads 0 on every row.	4	100%
company_profiles	similarCompanyNames	array<string>	Distinct names of companies the source lists as similar to this one. Empty array when none.	['Google','Microsoft']	4.8%
company_profiles	sourceEntityType	string	Source-reported legal / organizational structure of the company (normalized). The source-reported companion to the classifier-inferred inferredEntityType.	Public Company, Privately Held, Nonprofit, Government Agency	40.0%
company_profiles	stockExchange	string	Exchange the linked ticker trades on, one spelling per venue (over-the-counter markets read OTC, Xetra reads XETRA, NYSE American reads NYSE AMERICAN). Venues are never merged. NULL when there is no ticker link or the link records no venue.	NASDAQ	0.1%
company_profiles	tickerMatchConfidence	float	Confidence of the company-to-ticker link (0-1). Links below 0.85 are not shipped.	0.97	0.2%
company_profiles	tickerMatchMethod	string	How the link was established: curated (a reviewed link), website (the employer's own domain matched the issuer's), name_exact, registrant_name (the issuer's registered filer name), name_variant, name_fuzzy, exact_alias (a reviewed alias that links only the one name it lists), corporate_family (inherited from a corporate parent or sibling link), curated_franchise (a reviewed franchise link) or brand_operator (a franchisee or partner whose own name pairs a listed brand with an operator, such as "DWMJ Inc d/b/a McDonald's"). Not every method occurs in every release; the highest-precision subset is curated, website and name_exact.	website	0.2%
company_profiles	tickerParentName	string	Verified registered name of the listed entity that owns the ticker. For self / former_name rows this is the company's own registered name; for subsidiaries it is the parent's.	Apple Inc.	0.1%
company_profiles	tickerRelation	string	How the company relates to the linked ticker: self (the listed company), former_name (the listed company under an earlier name), subsidiary, division or acquisition (owned by the listed company), franchise (an independent franchisee operating under the listed company's brand, no equity link), partner (a contracted operator under the brand, such as a delivery or freight partner, no equity link), fund_portfolio (a company held by funds a listed manager runs) or bdc_holding (a portfolio investment of a listed business development company). NULL when there is no listed-company link.	self	0.2%
company_profiles	tickerSymbol	string	Stock ticker of the listed company this company resolves to (the company itself, or its listed parent or acquirer). Set only on the profiles linked to a listed company; NULL otherwise.	AAPL	0.2%
company_profiles	topEmployeeTitles	array<string>	Up to 10 most-frequent employee job titles observed across the company's source profiles (canonical-level topK). Empty array when none.	['Software Engineer','Program Manager']	0.3%
company_profiles	topRelatedCompanies	array<string>	Reserved for a later release; empty on every row of this release.	['audible','twitch']	0%
company_profiles	topRelatedCompaniesAvgSimilarity	array<float>	Reserved for a later release; empty on every row of this release.	[0.91,0.88]	0%
company_profiles	sourceRecordsCount	integer	Number of source records that name this company under any of the spellings that resolve to it, counted across job postings, company-database profile records and employee-review records. Not a posting count: use lifetimePostingsCount / lifetimeUniquePostingsCount for hiring volume.	92341	100%
company_contacts	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.	banfield pet hospital	100%
company_contacts	contactEmails	array<string>	De-duplicated recruiting and contact email addresses found across the company's postings. The per-posting contacts behind this roll-up ship in Job Postings Data (jobs_contacts).	['careers@example-vet.com','hr@example-vet.com']	61.3%
company_contacts	contactPhones	array<string>	Deduplicated contact phone numbers aggregated across all of the company's postings.	['+1-555-0142']	45.1%
company_contacts	contactURLs	array<string>	De-duplicated application and contact URLs found across the company's postings.	['https://example-vet.com/careers']	30.9%
company_contacts	postingsWithContactsCount	integer	Number of the company's postings that carried at least one extracted contact (support for the aggregation).	37	100%
company_contacts	contactLastSeenDate	timestamp	Most recent last-seen date across the postings that contributed contacts (freshness signal for the contact set).	2026-06-28 00:00:00.000	100%
company_hiring_activity	companyPostedSalaryTier	string	Coarse tier of the month's average annual salary across its postings, posted or predicted: high (100K USD or more), medium (60K to 100K USD), low (under 60K USD). NULL when no posting of the month has a salary.	high	99.4%
company_hiring_activity	companyStatus	string	Ownership class of the company, always populated: public (the listed company itself), subsidiary_of_public (a business it owns), franchisee_of_public or partner_of_public (a franchisee or contracted operator under its brand, no equity link), portfolio_company_of_public (held by funds a listed manager runs), government, nonprofit, educational, private, or unknown. On a row with ticker credit it follows the credit.	public	100%
company_hiring_activity	distinctSOCCount	integer	Distinct US SOC codes appearing in this month.	14	100%
company_hiring_activity	distinctStatesCount	integer	Distinct US states the company posted roles in.	23	100%
company_hiring_activity	growth90DaysVs365Days	float	Ratio of the last 3 months' unique postings to the trailing 12 months' average quarter, both counted only over the sources live in every month of the 12-month window (like-for-like), so the panel's first 11 months carry no value. Near 1.0 is steady hiring, above 1.5 acceleration, below 0.7 a slowdown. NULL when the trailing 12-month total is below 20 unique postings, the company has fewer than 6 months of observed activity, or any month in the window is partial.	1.68	37.4%
company_hiring_activity	inferredEntityType	string	Hiring role of the company, always populated: real_employer, staffing_agency, job_board_aggregator, not_a_company or unknown. A role ships only at 0.70 confidence or more, so unknown means not classified.	real_employer	100%
company_hiring_activity	isLowVolume	integer	1 if the company had fewer than 10 postings this month; its shares (pay stated, remote) rest on few postings.	0	100%
company_hiring_activity	isPartialMonth	integer	1 if the month is cut short by the data boundary, so its counts are not comparable to a whole month: the panel's first month when collection starts after the 1st, and the month holding the data cutoff. Leave these rows out of a trend; the counts are real but incomplete. The growth columns are already NULL wherever a comparison touches one.	0	100%
company_hiring_activity	medianDaysOpen	float	Reserved for a later release; empty on every row of this release.	21.5	0%
company_hiring_activity	medianDaysOpenPostingsCount	integer	Reserved for a later release; empty on every row of this release.	36	0%
company_hiring_activity	monthOverMonthGrowth	float	Like-for-like month-over-month growth of unique postings: this month's unique postings from the sources live in both months, over the previous month's from the same sources, minus 1. A source starting or stopping therefore does not move it. NULL when that prior count is under 10 unique postings, this month has fewer than 5, or either month is partial.	0.142	21.6%
company_hiring_activity	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.	Walmart	100%
company_hiring_activity	postingsWithAnyPostedSalaryCount	integer	Postings whose text or the source site's structured salary field states pay (salarySource description or source_field, value > 0), among the postings whose text was observed. Includes structured pay that may be a site's own unflagged estimate (see postingsWithSalaryCount), so compare it across time only within one source (company_hiring_activity_by_source).	312	100%
company_hiring_activity	postingsWithClearanceCount	integer	Postings whose description mentions a security clearance requirement, among postings whose text was observed (share over postingsWithContentCount).	5	100%
company_hiring_activity	postingsWithContentCount	integer	Postings in this row whose text (description) was observed: the denominator of every share above. A posting whose text was never collected cannot show pay, remote work, visa or clearance language, so it is in totalPostingsCount and in no share; dividing by totalPostingsCount would read a gap in description collection as a change in hiring.	455	100%
company_hiring_activity	postingsWithRemoteCount	integer	Flagged as remote work, among the postings whose text was observed (share over postingsWithContentCount).	89	100%
company_hiring_activity	postingsWithSalaryCount	integer	Postings whose text states the pay (salarySource = 'description' and a resolved annualSalaryAverageUSD above 0), among postings whose text was observed. A source site's structured salary field is not counted here, because some sites fill it with their own unflagged estimate; it is in postingsWithAnyPostedSalaryCount. Predicted salaries and flagged source estimates are in neither. Share = postingsWithSalaryCount / postingsWithContentCount.	241	100%
company_hiring_activity	postingsWithVisaSponsorshipCount	integer	Postings whose description mentions visa sponsorship, whether offering or declining it, among postings whose text was observed (share over postingsWithContentCount).	23	100%
company_hiring_activity	snapshotMonth	integer	The New York calendar month of the postings' jobDate (the stated posting date; else the first collection day, less any relative "N days ago" the posting shows; else the first collection day), YYYYMM.	202603	100%
company_hiring_activity	topSOCCounts	array<integer>	Parallel posting counts for topSOCs.	[152, 78]	100%
company_hiring_activity	topSOCs	array<string>	Top 10 occupation codes ranked by posting count.	['15-1252', '15-1132']	100%
company_hiring_activity	topStateCounts	array<integer>	Parallel posting counts for topStates.	[201, 89]	99.7%
company_hiring_activity	topStates	array<string>	Top 10 states ranked by posting count.	['WA', 'CA']	99.7%
company_hiring_activity	totalPostingsCount	integer	Postings attributed to this company in this month, one per posting row before near-duplicate collapse, so a job relisted under many source keys counts each time; use totalUniquePostingsCount for hiring demand. A company-month with fewer than 5 rows has no row here unless a listed security's owned counts credit it.	487	100%
company_hiring_activity	totalPostingsCountKnownAtPeriodEnd	integer	totalPostingsCount over only the postings first seen by the month's last day (New York time): what a reader holding the data when the month closed would have counted. It does not change after the month closes.	67965	100%
company_hiring_activity	totalUniquePostingsCount	integer	De-duplicated monthly posting count: unique postings only (isUniqueJobPosting = 1).	402	100%
company_hiring_activity	totalUniquePostingsCountKnownAtPeriodEnd	integer	totalUniquePostingsCount over only the postings first seen by the month's last day (New York time): what a reader holding the data when the month closed would have counted. It does not change after the month closes.	29521	100%
company_hiring_activity	yearOverYearGrowth	float	Like-for-like year-over-year growth of unique postings: the same rule against the same month 12 months earlier, over the sources live in both. NULL when the year-ago count is under 10 unique postings, this month has fewer than 5, or either month is partial.	-0.081	12.0%
company_hiring_activity_by_source	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.		100%
company_hiring_activity_by_source	snapshotMonth	integer	The New York calendar month of the postings' jobDate (the stated posting date; else the first collection day, less any relative "N days ago" the posting shows; else the first collection day), YYYYMM.		100%
company_hiring_activity_by_source	sourceWebsite	string	The source the postings came from. The rows of one company-month sum exactly to that month's counts in company_hiring_activity.		100%
company_hiring_activity_by_source	companyStatus	string	Ownership class of the company, always populated: public (the listed company itself), subsidiary_of_public (a business it owns), franchisee_of_public or partner_of_public (a franchisee or contracted operator under its brand, no equity link), portfolio_company_of_public (held by funds a listed manager runs), government, nonprofit, educational, private, or unknown. On a row with ticker credit it follows the credit.		100%
company_hiring_activity_by_source	postingsWithAnyPostedSalaryCount	integer	Postings whose text or the source site's structured salary field states pay (salarySource description or source_field, value > 0), among the postings whose text was observed. Includes structured pay that may be a site's own unflagged estimate (see postingsWithSalaryCount), so compare it across time only within one source (company_hiring_activity_by_source).		100%
company_hiring_activity_by_source	postingsWithClearanceCount	integer	Postings whose description mentions a security clearance requirement, among postings whose text was observed (share over postingsWithContentCount).		100%
company_hiring_activity_by_source	postingsWithContentCount	integer	Postings in this row whose text (description) was observed: the denominator of every share above. A posting whose text was never collected cannot show pay, remote work, visa or clearance language, so it is in totalPostingsCount and in no share; dividing by totalPostingsCount would read a gap in description collection as a change in hiring.		100%
company_hiring_activity_by_source	postingsWithRemoteCount	integer	Flagged as remote work, among the postings whose text was observed (share over postingsWithContentCount).		100%
company_hiring_activity_by_source	postingsWithSalaryCount	integer	Postings whose text states the pay (salarySource = 'description' and a resolved annualSalaryAverageUSD above 0), among postings whose text was observed. A source site's structured salary field is not counted here, because some sites fill it with their own unflagged estimate; it is in postingsWithAnyPostedSalaryCount. Predicted salaries and flagged source estimates are in neither. Share = postingsWithSalaryCount / postingsWithContentCount.		100%
company_hiring_activity_by_source	postingsWithVisaSponsorshipCount	integer	Postings whose description mentions visa sponsorship, whether offering or declining it, among postings whose text was observed (share over postingsWithContentCount).		100%
company_hiring_activity_by_source	totalPostingsCount	integer	Postings attributed to this company in this month, one per posting row before near-duplicate collapse, so a job relisted under many source keys counts each time; use totalUniquePostingsCount for hiring demand. A company-month with fewer than 5 rows has no row here unless a listed security's owned counts credit it.		100%
company_hiring_activity_by_source	totalPostingsCountKnownAtPeriodEnd	integer	totalPostingsCount over only the postings first seen by the month's last day (New York time): what a reader holding the data when the month closed would have counted. It does not change after the month closes.		100%
company_hiring_activity_by_source	totalUniquePostingsCount	integer	De-duplicated monthly posting count: unique postings only (isUniqueJobPosting = 1).		100%
company_hiring_activity_by_source	totalUniquePostingsCountKnownAtPeriodEnd	integer	totalUniquePostingsCount over only the postings first seen by the month's last day (New York time): what a reader holding the data when the month closed would have counted. It does not change after the month closes.		100%
company_hiring_activity_vintages	companyPostedSalaryTier	string	Coarse tier of the month's average annual salary across its postings, posted or predicted: high (100K USD or more), medium (60K to 100K USD), low (under 60K USD). NULL when no posting of the month has a salary.		99.4%
company_hiring_activity_vintages	companyStatus	string	Ownership class of the company, always populated: public (the listed company itself), subsidiary_of_public (a business it owns), franchisee_of_public or partner_of_public (a franchisee or contracted operator under its brand, no equity link), portfolio_company_of_public (held by funds a listed manager runs), government, nonprofit, educational, private, or unknown. On a row with ticker credit it follows the credit.		100%
company_hiring_activity_vintages	distinctSOCCount	integer	Distinct US SOC codes appearing in this month.		100%
company_hiring_activity_vintages	distinctStatesCount	integer	Distinct US states the company posted roles in.		100%
company_hiring_activity_vintages	growth90DaysVs365Days	float	Ratio of the last 3 months' unique postings to the trailing 12 months' average quarter, both counted only over the sources live in every month of the 12-month window (like-for-like), so the panel's first 11 months carry no value. Near 1.0 is steady hiring, above 1.5 acceleration, below 0.7 a slowdown. NULL when the trailing 12-month total is below 20 unique postings, the company has fewer than 6 months of observed activity, or any month in the window is partial.		38.0%
company_hiring_activity_vintages	inferredEntityType	string	Hiring role of the company, always populated: real_employer, staffing_agency, job_board_aggregator, not_a_company or unknown. A role ships only at 0.70 confidence or more, so unknown means not classified.		100%
company_hiring_activity_vintages	isLowVolume	integer	1 if the company had fewer than 10 postings this month; its shares (pay stated, remote) rest on few postings.		100%
company_hiring_activity_vintages	isPartialMonth	integer	1 if the month is cut short by the data boundary, so its counts are not comparable to a whole month: the panel's first month when collection starts after the 1st, and the month holding the data cutoff. Leave these rows out of a trend; the counts are real but incomplete. The growth columns are already NULL wherever a comparison touches one.		100%
company_hiring_activity_vintages	medianDaysOpen	float	Reserved for a later release; empty on every row of this release.		0%
company_hiring_activity_vintages	medianDaysOpenPostingsCount	integer	Reserved for a later release; empty on every row of this release.		0%
company_hiring_activity_vintages	monthOverMonthGrowth	float	Like-for-like month-over-month growth of unique postings: this month's unique postings from the sources live in both months, over the previous month's from the same sources, minus 1. A source starting or stopping therefore does not move it. NULL when that prior count is under 10 unique postings, this month has fewer than 5, or either month is partial.		22.0%
company_hiring_activity_vintages	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.		100%
company_hiring_activity_vintages	postingsWithAnyPostedSalaryCount	integer	Postings whose text or the source site's structured salary field states pay (salarySource description or source_field, value > 0), among the postings whose text was observed. Includes structured pay that may be a site's own unflagged estimate (see postingsWithSalaryCount), so compare it across time only within one source (company_hiring_activity_by_source).		100%
company_hiring_activity_vintages	postingsWithClearanceCount	integer	Postings whose description mentions a security clearance requirement, among postings whose text was observed (share over postingsWithContentCount).		100%
company_hiring_activity_vintages	postingsWithContentCount	integer	Postings in this row whose text (description) was observed: the denominator of every share above. A posting whose text was never collected cannot show pay, remote work, visa or clearance language, so it is in totalPostingsCount and in no share; dividing by totalPostingsCount would read a gap in description collection as a change in hiring.		100%
company_hiring_activity_vintages	postingsWithRemoteCount	integer	Flagged as remote work, among the postings whose text was observed (share over postingsWithContentCount).		100%
company_hiring_activity_vintages	postingsWithSalaryCount	integer	Postings whose text states the pay (salarySource = 'description' and a resolved annualSalaryAverageUSD above 0), among postings whose text was observed. A source site's structured salary field is not counted here, because some sites fill it with their own unflagged estimate; it is in postingsWithAnyPostedSalaryCount. Predicted salaries and flagged source estimates are in neither. Share = postingsWithSalaryCount / postingsWithContentCount.		100%
company_hiring_activity_vintages	postingsWithVisaSponsorshipCount	integer	Postings whose description mentions visa sponsorship, whether offering or declining it, among postings whose text was observed (share over postingsWithContentCount).		100%
company_hiring_activity_vintages	snapshotMonth	integer	The New York calendar month of the postings' jobDate (the stated posting date; else the first collection day, less any relative "N days ago" the posting shows; else the first collection day), YYYYMM.		100%
company_hiring_activity_vintages	topSOCCounts	array<integer>	Parallel posting counts for topSOCs.		100%
company_hiring_activity_vintages	topSOCs	array<string>	Top 10 occupation codes ranked by posting count.		100%
company_hiring_activity_vintages	topStateCounts	array<integer>	Parallel posting counts for topStates.		99.7%
company_hiring_activity_vintages	topStates	array<string>	Top 10 states ranked by posting count.		99.7%
company_hiring_activity_vintages	totalPostingsCount	integer	Postings attributed to this company in this month, one per posting row before near-duplicate collapse, so a job relisted under many source keys counts each time; use totalUniquePostingsCount for hiring demand. A company-month with fewer than 5 rows has no row here unless a listed security's owned counts credit it.		100%
company_hiring_activity_vintages	totalPostingsCountKnownAtPeriodEnd	integer	totalPostingsCount over only the postings first seen by the month's last day (New York time): what a reader holding the data when the month closed would have counted. It does not change after the month closes.		100%
company_hiring_activity_vintages	totalUniquePostingsCount	integer	De-duplicated monthly posting count: unique postings only (isUniqueJobPosting = 1).		100%
company_hiring_activity_vintages	totalUniquePostingsCountKnownAtPeriodEnd	integer	totalUniquePostingsCount over only the postings first seen by the month's last day (New York time): what a reader holding the data when the month closed would have counted. It does not change after the month closes.		100%
company_hiring_activity_vintages	yearOverYearGrowth	float	Like-for-like year-over-year growth of unique postings: the same rule against the same month 12 months earlier, over the sources live in both. NULL when the year-ago count is under 10 unique postings, this month has fewer than 5, or either month is partial.		12.2%
company_hiring_activity_vintages	vintage	string	The release that delivered this row. Together with the file's key it identifies one as-known value: a period a later release restated keeps its earlier row beside the new one.		100%
company_hiring_activity_vintages	vintageRecordedAt	timestamp	When this vintage row was first recorded (UTC). Rows are appended when a delivery first sees a release's value and are never rewritten.		100%
company_hiring_activity_vintages	vintageRetracted	integer	1 on the row recorded when a release no longer carried a period an earlier release had carried (the row repeats the last values, so the retraction is visible); 0 otherwise.	0	100%
company_technologies	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.		100%
company_technologies	technology	string	A named software product, platform or programming language (for example Snowflake, Kubernetes, Python, Salesforce); business, soft and trade skills, equipment and certifications are not included. Found by matching product names in posting text; for names that are also ordinary words, require uniquePostings of 2 or more when precision matters.	Python	100%
company_technologies	firstSeenDate	timestamp	Earliest first-seen date among the company's unique postings naming the technology.		100%
company_technologies	lastSeenDate	timestamp	Latest last-seen date among the company's unique postings naming the technology.		100%
company_technologies	uniquePostings	integer	The company's unique postings (one per duplicate group) naming the technology, each counted once.		100%
company_technologies_monthly	technology	string	As in company_technologies.	Python	100%
company_technologies_monthly	snapshotMonth	integer	YYYYMM of the postings' first sighting (New York calendar).		100%
company_technologies_monthly	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.		100%
company_technologies_monthly	firstSeenDate	timestamp	Earliest first-seen date among the month's unique postings naming the technology.		100%
company_technologies_monthly	lastSeenDate	timestamp	Latest last-seen date among the month's unique postings naming the technology.		100%
company_technologies_monthly	uniquePostings	integer	The company's unique postings first seen in the month that name the technology.		100%
company_hiring_events	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.		100%
company_hiring_events	eventType	string	surge (the month's unique postings are at least twice the mean of the 3 months before and at least 10, counting only sources live in all four months; never the newest month), new_location (the company's first posting in a US state, or in a new city of a state it already posts in), or new_technology (the company's first posting naming a technology). new_location and new_technology need six months of history with the company and with the source, so a new source or new coverage is not read as expansion.	surge	100%
company_hiring_events	eventDate	timestamp	First-seen time of the evidence posting, or for a surge the first instant after the month (New York midnight).		100%
company_hiring_events	baselinePostings	float	For a surge, the mean monthly unique postings of the 3 months before. NULL for other event types.		1.7%
company_hiring_events	companyFirstSeenDate	timestamp	First-seen date of the company's earliest posting, the start of the history the event rules require.		100%
company_hiring_events	eventPostings	integer	Unique postings behind the event: the month's count for a surge, the postings in the new place or naming the new technology that month otherwise.		100%
company_hiring_events	eventRule	string	The rule that fired, in words, so every event can be audited.		100%
company_hiring_events	finalCity	string	City of a new_location event at city level. NULL for other events.		58.8%
company_hiring_events	finalState	string	US state or territory of a new_location event; a handful carry a region outside the US as the posting wrote it. NULL for other events.		58.9%
company_hiring_events	jobID	string	Posting key of the event's evidence posting, joining to the job postings files; NULL for surge events.		98.3%
company_hiring_events	locationLevel	string	state or city for a new_location event. NULL for other events.		58.9%
company_hiring_events	snapshotMonth	integer	YYYYMM (New York calendar) of the event's first sighting.		100%
company_hiring_events	technology	string	The technology of a new_technology event. NULL for other events.		39.4%
company_headcount_monthly	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.		100%
company_headcount_monthly	snapshotMonth	integer	Calendar month, YYYYMM.	202607	100%
company_headcount_monthly	companyLinkedInURL	string	The LinkedIn company page the employee count was read from.		100%
company_headcount_monthly	employeeCount	integer	The employee count the page showed: the month's latest observation, or the latest earlier one carried forward.		100%
company_headcount_monthly	employeeCountObservedDate	date	The date employeeCount was observed.		100%
company_headcount_monthly	monthsSinceObservation	integer	0 when observed in this month, else whole months since employeeCountObservedDate.		100%
company_headcount_monthly	isCarriedForward	integer	1 when monthsSinceObservation > 0.		100%
company_headcount_monthly	observationsInMonthCount	integer	Distinct observation dates in this month (0 when carried forward).		100%
company_funding_events	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.		-
company_funding_events	fundingFilingDate	date	Date the filing was made.		-
company_funding_events	fundingFilingID	string	The filing's identifier, unique per filing.		-
company_funding_events	isFundingAmendment	integer	1 for an amendment of an earlier filing, 0 for a new notice; NULL when not stated.		-
company_funding_events	fundingFirstSaleDate	date	Date of the offering's first sale; NULL when it had not yet occurred or was not stated.		-
company_funding_events	fundingOfferingAmountUSD	float	Total offering amount in USD; NULL when the offering amount is indefinite or not stated.		-
company_funding_events	isFundingOfferingAmountIndefinite	integer	1 when the filing states an indefinite offering amount.		-
company_funding_events	fundingAmountSoldUSD	float	Amount sold as of the filing, in USD; NULL when not stated.		-
company_funding_events	fundingAmountRemainingUSD	float	Amount remaining to be sold, in USD; NULL when indefinite or not stated.		-
company_funding_events	fundingInvestorCount	integer	Number of investors who had already invested.		-
company_funding_events	fundingIndustryGroup	string	Industry group as filed.		-
company_funding_events	fundingSecurityTypes	array<string>	Securities offered, any of equity, debt, pooled_investment_fund; empty when the filing reports none of them.		-
company_funding_events	fundingRevenueRange	string	The issuer's revenue range as filed, including Decline to Disclose; empty when not stated.		-

Public Company Hiring Data: 60 fields, 293,064 rows (catalog 8.3, measured 6 Oct 2026)

Module	Field Name	Type	Description	Example	Coverage
ticker_hiring_activity	tickerSymbol	string	The posting's ticker, point in time.	ADEN	100%
ticker_hiring_activity	stockExchange	string, nullable	The venue: NYSE, NASDAQ, NYSE AMERICAN, OTHER (NYSE Arca, Cboe BZX and IEX) or OTC (over-the-counter quotations) for US listings; the exchange's name outside the US. Two securities can hold one symbol in one period, so key on tickerTenancyID as well, never on the symbol alone. NULL keeps postings whose link names no venue, or several, on a row of their own.	NASDAQ	85.0%
ticker_hiring_activity	tickerTenancyID	string	Which security held the symbol: an opaque 16-character label, one per security's stay on a symbol. Every count is keyed on it beside the symbol and venue, so a symbol that passed from one company to another is two series. Never NULL or empty.	e98003753a5368b8	100%
ticker_hiring_activity	snapshotMonth	integer	YYYYMM of jobDate (New York calendar).	202606	100%
ticker_hiring_activity	grossPostings	integer	Owned postings: the listed company and what it owns (tickerRelation self, former_name, subsidiary, division or acquisition), counted once per source listing, so a source relisting one job under new keys adds it once, in the month of their posting date (jobDate, New York calendar).	1	100%
ticker_hiring_activity	grossPostingsInclFranchise	integer	grossPostings plus the postings of franchisees and brand partners (tickerRelation franchise or partner: a brand licence or service contract, no equity link), in the month of their posting date (jobDate, New York calendar).	1	100%
ticker_hiring_activity	grossPostingsInclFranchiseKnownAtPeriodEnd	integer	grossPostingsInclFranchise counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity	grossPostingsKnownAtPeriodEnd	integer	grossPostings counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity	isPartialMonth	integer	1 on a month cut short by the data boundary.	0	100%
ticker_hiring_activity	marketCapTier	string, nullable	large, mid, small or micro, from publicFloatUSD. NULL where no float applies.	small	58.4%
ticker_hiring_activity	medianDaysOpen	float, nullable	Reserved for a later release; empty on every row of this release.		0%
ticker_hiring_activity	medianDaysOpenPostingsCount	integer, nullable	Reserved for a later release; empty on every row of this release.		0%
ticker_hiring_activity	publicFloatAsOf	date, nullable	The date publicFloatUSD is measured at: the last business day of the issuer's second fiscal quarter.	2023-06-30	58.4%
ticker_hiring_activity	publicFloatUSD	integer, nullable	The issuer's SEC public float in USD, from the latest annual report (10-K) filed on or before the first day of the month. NULL where no float applies.	1600000000	58.4%
ticker_hiring_activity	tickerCategory	string	What kind of business the ticker is: staffing, job_board, marketplace, franchisor or operating. Never NULL.	operating	100%
ticker_hiring_activity	uniquePostings	integer	grossPostings restricted to unique postings (isUniqueJobPosting = 1): each group of duplicates counts once, in the month of their posting date (jobDate, New York calendar).	1	100%
ticker_hiring_activity	uniquePostingsInclFranchise	integer	grossPostingsInclFranchise restricted to unique postings (isUniqueJobPosting = 1), in the month of their posting date (jobDate, New York calendar).	1	100%
ticker_hiring_activity	uniquePostingsInclFranchiseKnownAtPeriodEnd	integer	uniquePostingsInclFranchise counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity	uniquePostingsKnownAtPeriodEnd	integer	uniquePostings counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity	uniquePostingsLikeForLikeIndex	float, nullable	Chain-linked level index of uniquePostings over the sources live in both months, 100 in its segment's base month. A month with fewer than 20 matched unique postings on either side starts a new segment, and two index values compare only inside one segment.	100	98.3%
ticker_hiring_activity	uniquePostingsLikeForLikeIndexBaseMonth	integer, nullable	YYYYMM of the month this row's index level is 100 at. Two index values compare only when they share it. NULL where the index is NULL.	202201	98.3%
ticker_hiring_activity_weekly	tickerSymbol	string	The posting's ticker, point in time.	ADEN	100%
ticker_hiring_activity_weekly	stockExchange	string, nullable	The venue: NYSE, NASDAQ, NYSE AMERICAN, OTHER (NYSE Arca, Cboe BZX and IEX) or OTC (over-the-counter quotations) for US listings; the exchange's name outside the US. Two securities can hold one symbol in one period, so key on tickerTenancyID as well, never on the symbol alone. NULL keeps postings whose link names no venue, or several, on a row of their own.	NASDAQ	84.6%
ticker_hiring_activity_weekly	tickerTenancyID	string	Which security held the symbol: an opaque 16-character label, one per security's stay on a symbol. Every count is keyed on it beside the symbol and venue, so a symbol that passed from one company to another is two series. Never NULL or empty.	68f3966ab6a7a8d9	100%
ticker_hiring_activity_weekly	snapshotWeek	date	The Monday of the ISO week (New York calendar) of the posting's jobDate, the monthly file's counting date. A week that spans two months counts once here and in both months of the monthly file, so weeks do not sum to months.	2026-04-13	100%
ticker_hiring_activity_weekly	grossPostings	integer	Owned postings: the listed company and what it owns (tickerRelation self, former_name, subsidiary, division or acquisition), counted once per source listing, so a source relisting one job under new keys adds it once, in the ISO week (Monday start, New York calendar) of their posting date.	1	100%
ticker_hiring_activity_weekly	grossPostingsInclFranchise	integer	grossPostings plus the postings of franchisees and brand partners (tickerRelation franchise or partner: a brand licence or service contract, no equity link), in the ISO week (Monday start, New York calendar) of their posting date.	1	100%
ticker_hiring_activity_weekly	grossPostingsInclFranchiseKnownAtPeriodEnd	integer	grossPostingsInclFranchise counting only the postings first seen (New York day) by the last day of the week: the value as it was knowable when the week closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_weekly	grossPostingsKnownAtPeriodEnd	integer	grossPostings counting only the postings first seen (New York day) by the last day of the week: the value as it was knowable when the week closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_weekly	isPartialWeek	integer	1 on the first week when its first posting is not on a Monday, and on the newest week when its last posting is not on a Sunday.	0	100%
ticker_hiring_activity_weekly	knowableAt	datetime (UTC), nullable	When the row became knowable (UTC): the latest time our system first recorded any posting the row counts, so it never dates a count earlier than it was knowable. On a first release, rows carry the time the release was produced; for historical as-of work, use the KnownAtPeriodEnd columns and the daily ...Arrived counts.	2026-10-04 17:20:25.246	100%
ticker_hiring_activity_weekly	marketCapTier	string, nullable	large, mid, small or micro, from publicFloatUSD. NULL where no float applies.	small	59.4%
ticker_hiring_activity_weekly	publicFloatAsOf	date, nullable	The date publicFloatUSD is measured at: the last business day of the issuer's second fiscal quarter.	2023-06-30	59.4%
ticker_hiring_activity_weekly	publicFloatUSD	integer, nullable	The issuer's SEC public float in USD, from the latest annual report (10-K) filed on or before snapshotWeek (the week's Monday). NULL where no float applies.	1700000000	59.4%
ticker_hiring_activity_weekly	tickerCategory	string	What kind of business the ticker is: staffing, job_board, marketplace, franchisor or operating. Never NULL.	operating	100%
ticker_hiring_activity_weekly	uniquePostings	integer	grossPostings restricted to unique postings (isUniqueJobPosting = 1): each group of duplicates counts once, in the ISO week (Monday start, New York calendar) of their posting date.	1	100%
ticker_hiring_activity_weekly	uniquePostingsInclFranchise	integer	grossPostingsInclFranchise restricted to unique postings (isUniqueJobPosting = 1), in the ISO week (Monday start, New York calendar) of their posting date.	1	100%
ticker_hiring_activity_weekly	uniquePostingsInclFranchiseKnownAtPeriodEnd	integer	uniquePostingsInclFranchise counting only the postings first seen (New York day) by the last day of the week: the value as it was knowable when the week closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_weekly	uniquePostingsKnownAtPeriodEnd	integer	uniquePostings counting only the postings first seen (New York day) by the last day of the week: the value as it was knowable when the week closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_breakdown	tickerSymbol	string	The posting's ticker, point in time.	CMPGY	100%
ticker_hiring_activity_breakdown	stockExchange	string, nullable	The venue: NYSE, NASDAQ, NYSE AMERICAN, OTHER (NYSE Arca, Cboe BZX and IEX) or OTC (over-the-counter quotations) for US listings; the exchange's name outside the US. Two securities can hold one symbol in one period, so key on tickerTenancyID as well, never on the symbol alone. NULL keeps postings whose link names no venue, or several, on a row of their own.	NYSE	85.3%
ticker_hiring_activity_breakdown	tickerTenancyID	string	Which security held the symbol: an opaque 16-character label, one per security's stay on a symbol. Every count is keyed on it beside the symbol and venue, so a symbol that passed from one company to another is two series. Never NULL or empty.	081ee22ad45120c3	100%
ticker_hiring_activity_breakdown	snapshotMonth	integer	YYYYMM of jobDate (New York calendar).	202606	100%
ticker_hiring_activity_breakdown	breakdownDimension	string	socMajorGroup (the SOC 2018 major group of socCode, as NN-0000), state (finalState), seniority (seniority) or workArrangement (remoteStatus).	state	100%
ticker_hiring_activity_breakdown	breakdownValue	string, nullable	The value. NULL counts the postings with no value for the dimension.	On-Site	98.7%
ticker_hiring_activity_breakdown	grossPostings	integer	Owned postings: the listed company and what it owns (tickerRelation self, former_name, subsidiary, division or acquisition), counted once per source listing, so a source relisting one job under new keys adds it once, in the month, with this breakdownValue.	1	100%
ticker_hiring_activity_breakdown	grossPostingsInclFranchise	integer	grossPostings plus the postings of franchisees and brand partners (tickerRelation franchise or partner: a brand licence or service contract, no equity link), in the month, with this breakdownValue.	1	100%
ticker_hiring_activity_breakdown	grossPostingsInclFranchiseKnownAtPeriodEnd	integer	grossPostingsInclFranchise counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_breakdown	grossPostingsKnownAtPeriodEnd	integer	grossPostings counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_breakdown	isPartialMonth	integer	ticker_hiring_activity's flag for the month.	0	100%
ticker_hiring_activity_breakdown	knowableAt	datetime (UTC), nullable	When the row became knowable (UTC): the latest time our system first recorded any posting the row counts, so it never dates a count earlier than it was knowable. On a first release, rows carry the time the release was produced; for historical as-of work, use the KnownAtPeriodEnd columns and the daily ...Arrived counts.	2026-10-04 17:20:25.246	100%
ticker_hiring_activity_breakdown	marketCapTier	string, nullable	ticker_hiring_activity's values for the month.	large	64.4%
ticker_hiring_activity_breakdown	publicFloatAsOf	date, nullable	ticker_hiring_activity's values for the month.	2023-06-30	64.4%
ticker_hiring_activity_breakdown	publicFloatUSD	integer, nullable	ticker_hiring_activity's values for the month.	3000000000	64.4%
ticker_hiring_activity_breakdown	tickerCategory	string	What kind of business the ticker is: staffing, job_board, marketplace, franchisor or operating. Never NULL.	operating	100%
ticker_hiring_activity_breakdown	uniquePostings	integer	grossPostings restricted to unique postings (isUniqueJobPosting = 1): each group of duplicates counts once, in the month, with this breakdownValue.	1	100%
ticker_hiring_activity_breakdown	uniquePostingsInclFranchise	integer	grossPostingsInclFranchise restricted to unique postings (isUniqueJobPosting = 1), in the month, with this breakdownValue.	1	100%
ticker_hiring_activity_breakdown	uniquePostingsInclFranchiseKnownAtPeriodEnd	integer	uniquePostingsInclFranchise counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_breakdown	uniquePostingsKnownAtPeriodEnd	integer	uniquePostings counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_by_source	tickerSymbol	string	The posting's ticker, point in time.	ADEN	100%
ticker_hiring_activity_by_source	stockExchange	string, nullable	The venue: NYSE, NASDAQ, NYSE AMERICAN, OTHER (NYSE Arca, Cboe BZX and IEX) or OTC (over-the-counter quotations) for US listings; the exchange's name outside the US. Two securities can hold one symbol in one period, so key on tickerTenancyID as well, never on the symbol alone. NULL keeps postings whose link names no venue, or several, on a row of their own.	NYSE	85.7%
ticker_hiring_activity_by_source	tickerTenancyID	string	Which security held the symbol: an opaque 16-character label, one per security's stay on a symbol. Every count is keyed on it beside the symbol and venue, so a symbol that passed from one company to another is two series. Never NULL or empty.	02c7ec0490ff6f83	100%
ticker_hiring_activity_by_source	snapshotMonth	integer	YYYYMM of jobDate (New York calendar).	202401	100%
ticker_hiring_activity_by_source	sourceWebsite	string	The source, as on the job postings files.	indeed	100%
ticker_hiring_activity_by_source	grossPostings	integer	Owned postings: the listed company and what it owns (tickerRelation self, former_name, subsidiary, division or acquisition), counted once per source listing, so a source relisting one job under new keys adds it once, in the month, from this sourceWebsite.	1	100%
ticker_hiring_activity_by_source	grossPostingsInclFranchise	integer	grossPostings plus the postings of franchisees and brand partners (tickerRelation franchise or partner: a brand licence or service contract, no equity link), in the month, from this sourceWebsite.	1	100%
ticker_hiring_activity_by_source	grossPostingsInclFranchiseKnownAtPeriodEnd	integer	grossPostingsInclFranchise counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_by_source	grossPostingsKnownAtPeriodEnd	integer	grossPostings counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_by_source	isMatchedMoM	integer	1 when the source is live in this month and the month before, so this row counts toward like-for-like month-over-month growth; 0 otherwise.	1	100%
ticker_hiring_activity_by_source	isMatchedYoY	integer	1 when the source is live in this month and the same month a year earlier, so this row counts toward like-for-like year-over-year growth; 0 otherwise.	0	100%
ticker_hiring_activity_by_source	isPartialMonth	integer	ticker_hiring_activity's flag for the month.	0	100%
ticker_hiring_activity_by_source	knowableAt	datetime (UTC), nullable	When the row became knowable (UTC): the latest time our system first recorded any posting the row counts, so it never dates a count earlier than it was knowable. On a first release, rows carry the time the release was produced; for historical as-of work, use the KnownAtPeriodEnd columns and the daily ...Arrived counts.	2026-10-04 17:20:25.246	100%
ticker_hiring_activity_by_source	tickerCategory	string	What kind of business the ticker is: staffing, job_board, marketplace, franchisor or operating. Never NULL.	operating	100%
ticker_hiring_activity_by_source	uniquePostings	integer	grossPostings restricted to unique postings (isUniqueJobPosting = 1): each group of duplicates counts once, in the month, from this sourceWebsite.	1	100%
ticker_hiring_activity_by_source	uniquePostingsInclFranchise	integer	grossPostingsInclFranchise restricted to unique postings (isUniqueJobPosting = 1), in the month, from this sourceWebsite.	1	100%
ticker_hiring_activity_by_source	uniquePostingsInclFranchiseKnownAtPeriodEnd	integer	uniquePostingsInclFranchise counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_activity_by_source	uniquePostingsKnownAtPeriodEnd	integer	uniquePostings counting only the postings first seen (New York day) by the last day of the month: the value as it was knowable when the month closed. Point in time; it never gains a posting first seen later.	1	100%
ticker_hiring_daily	tickerSymbol	string	The posting's ticker, point in time.	ADRNY	100%
ticker_hiring_daily	stockExchange	string, nullable	The venue: NYSE, NASDAQ, NYSE AMERICAN, OTHER (NYSE Arca, Cboe BZX and IEX) or OTC (over-the-counter quotations) for US listings; the exchange's name outside the US. Two securities can hold one symbol in one period, so key on tickerTenancyID as well, never on the symbol alone. NULL keeps postings whose link names no venue, or several, on a row of their own.	NASDAQ	85.7%
ticker_hiring_daily	tickerTenancyID	string	Which security held the symbol: an opaque 16-character label, one per security's stay on a symbol. Key on it with tickerSymbol and stockExchange. Never NULL or empty.	3709bef09cf87198	100%
ticker_hiring_daily	snapshotDate	date	A New York calendar day.	2026-09-24	100%
ticker_hiring_daily	closedUniquePostings	integer, nullable	Reserved for a later release; empty on every row of this release.		0%
ticker_hiring_daily	grossPostings	integer	Owned postings: the listed company and what it owns (tickerRelation self, former_name, subsidiary, division or acquisition), counted once per source listing, so a source relisting one job under new keys adds it once, on this New York day of their posting date (jobDate).	0	100%
ticker_hiring_daily	grossPostingsArrived	integer	The same postings as grossPostings, counted on the New York day they were first seen rather than on their posting date: what reached us that day. Point in time; a day's unique arrivals are final when the day ends.	0	100%
ticker_hiring_daily	grossPostingsInclFranchise	integer	grossPostings plus the postings of franchisees and brand partners (tickerRelation franchise or partner: a brand licence or service contract, no equity link), on this New York day of their posting date (jobDate).	0	100%
ticker_hiring_daily	grossPostingsInclFranchiseArrived	integer	The same postings as grossPostingsInclFranchise, counted on the New York day they were first seen rather than on their posting date: what reached us that day. Point in time; a day's unique arrivals are final when the day ends.	0	100%
ticker_hiring_daily	grossPostingsInclFranchiseKnownAtPeriodEnd	integer	grossPostingsInclFranchise counting only the postings first seen (New York day) by the last day of the day: the value as it was knowable when the day closed. Point in time; it never gains a posting first seen later.	0	100%
ticker_hiring_daily	grossPostingsKnownAtPeriodEnd	integer	grossPostings counting only the postings first seen (New York day) by the last day of the day: the value as it was knowable when the day closed. Point in time; it never gains a posting first seen later.	0	100%
ticker_hiring_daily	isPartialDay	integer	1 on the newest days of the panel, whose collection may not be complete; every earlier day is closed.	0	100%
ticker_hiring_daily	knowableAt	datetime (UTC), nullable	When the row became knowable (UTC): the latest time our system first recorded any posting the row counts, or the end of the day on a day with no postings, so it never dates a count earlier than it was knowable. On a first release, rows that count postings carry the time the release was produced; for historical as-of work, use the KnownAtPeriodEnd columns and the ...Arrived counts.	2026-10-04 17:20:25.246	100%
ticker_hiring_daily	marketCapTier	string, nullable	large, mid, small or micro, from the publicFloatUSD in force on snapshotDate. NULL where no float applies.	micro	56.9%
ticker_hiring_daily	openUniquePostings	integer, nullable	Reserved for a later release; empty on every row of this release.		0%
ticker_hiring_daily	publicFloatAsOf	date, nullable	The date publicFloatUSD is measured at (the last business day of the issuer's second fiscal quarter), for the float in force on snapshotDate.	2023-06-30	56.9%
ticker_hiring_daily	publicFloatUSD	integer, nullable	The issuer's SEC public float in USD, from the latest annual report (10-K) filed on or before snapshotDate.	1100000000	56.9%
ticker_hiring_daily	sourcesCrawled	integer, nullable	Reserved for a later release; empty on every row of this release.		0%
ticker_hiring_daily	sourcesExpected	integer, nullable	Reserved for a later release; empty on every row of this release.		0%
ticker_hiring_daily	tickerCategory	string	What kind of business the ticker is: staffing, job_board, marketplace, franchisor or operating. Never NULL.	operating	100%
ticker_hiring_daily	uniquePostings	integer	grossPostings restricted to unique postings (isUniqueJobPosting = 1): each group of duplicates counts once, on this New York day of their posting date (jobDate).	0	100%
ticker_hiring_daily	uniquePostingsArrived	integer	The same postings as uniquePostings, counted on the New York day they were first seen rather than on their posting date: what reached us that day. Point in time; a day's unique arrivals are final when the day ends.	0	100%
ticker_hiring_daily	uniquePostingsInclFranchise	integer	grossPostingsInclFranchise restricted to unique postings (isUniqueJobPosting = 1), on this New York day of their posting date (jobDate).	0	100%
ticker_hiring_daily	uniquePostingsInclFranchiseArrived	integer	The same postings as uniquePostingsInclFranchise, counted on the New York day they were first seen rather than on their posting date: what reached us that day. Point in time; a day's unique arrivals are final when the day ends.	0	100%
ticker_hiring_daily	uniquePostingsInclFranchiseKnownAtPeriodEnd	integer	uniquePostingsInclFranchise counting only the postings first seen (New York day) by the last day of the day: the value as it was knowable when the day closed. Point in time; it never gains a posting first seen later.	0	100%
ticker_hiring_daily	uniquePostingsKnownAtPeriodEnd	integer	uniquePostings counting only the postings first seen (New York day) by the last day of the day: the value as it was knowable when the day closed. Point in time; it never gains a posting first seen later.	0	100%
ticker_hiring_daily	unmeasurableOpenPostings	integer, nullable	Reserved for a later release; empty on every row of this release.		0%
ticker_identifiers	tickerSymbol	string	Ticker symbol, as on ticker_hiring_activity. Join a series row on tickerSymbol plus stockExchange and its period date between validFrom and validTo.	CR	100%
ticker_identifiers	stockExchange	string, nullable	Venue, as on ticker_hiring_activity. A row whose venue is null carries no identifiers.	NASDAQ	88.7%
ticker_identifiers	tickerTenancyID	string	The series key's listing, as on every ticker file. A symbol two securities held in turn is two keys, each with only its own security's identifiers. Empty on a few rows, which match no series.	a484ce577c12f0d0	99.4%
ticker_identifiers	validFrom	date, nullable	First day the row holds, inclusive; NULL = from the start of the listing.	1997-12-31	73.3%
ticker_identifiers	validTo	date, nullable	Last day the row holds, inclusive; NULL = still current.	2023-03-28	24.4%
ticker_identifiers	compositeFIGI	string, nullable	The composite FIGI of the security that traded under the symbol over these days.	BBG017BXPZ85	61.6%
ticker_identifiers	shareClassFIGI	string, nullable	The share-class FIGI of that security.	BBG017BXPZ94	61.6%
ticker_identifiers	legalEntityIdentifier	string, nullable	The issuer's LEI, where it can be tied to that security.	549300EXNPSXEWUD3N21	50.1%
ticker_identifiers	issuerCik	string, nullable	The SEC filer (CIK) behind the listing, where every range of the key names the same one; NULL otherwise.	1789029	72.3%
ticker_identifiers	creditedTickerSymbol	string	The class whose series carry this symbol's postings: the most liquid class for a secondary share class (BRK.A to BRK.B, GOOG to GOOGL), else the symbol itself.	CR	100%
ticker_identifiers	secTickerSymbol	string, nullable	The symbol as the SEC's published ticker list spells it: a class suffix after a dash where these files use a dot (BRK.A as BRK-A). NULL on a venue outside the US.	CR	86.7%
company_ticker_history	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.	Verb Technology Company	100%
company_ticker_history	tickerSymbol	string	Stock ticker symbol in force during this listing window. Not unique across venues on its own; join on tickerSymbol plus stockExchange.	ANSS	100%
company_ticker_history	stockExchange	string	Exchange or venue of the listing (NYSE, NASDAQ, LSE, XETRA). Empty when no exchange was recorded for the link.	NASDAQ	95.0%
company_ticker_history	tickerRelation	string	How the company relates to the listed entity: self (the listed company), former_name (the listed company under an earlier name), subsidiary, division or acquisition (owned by the listed company), franchise (an independent franchisee operating under the listed company's brand, no equity link), partner (a contracted operator under the listed company's brand, no equity link), or fund_portfolio / bdc_holding (held by funds a listed manager runs, or by a business development company, not by the listed company itself).	self	100%
company_ticker_history	tickerParentName	string, nullable	Name of the listed entity that carries the ticker. NULL where the listed parent could not be named.	ANSYS INC	92.8%
company_ticker_history	isPubliclyTraded	integer	1 when the company is the listed entity itself (relation self or former_name); 0 when it is owned by a listed parent.	1	100%
company_ticker_history	isUSListing	integer, nullable	1 when the linked ticker trades on a US exchange, 0 for a venue outside the US, NULL when unknown (no exchange recorded, or an over-the-counter quotation, which stockExchange = 'OTC' flags in the ticker files). Describes tickerSymbol, not the primaryListingSymbol / primaryListingExchange pair.	1	94.1%
company_ticker_history	primaryListingSymbol	string, nullable	The company's home / ordinary-share listing symbol, carried beside tickerSymbol rather than instead of it. primaryListingSymbol != tickerSymbol confirms tickerSymbol is not the home listing (an ADR or another secondary line); it does not rule one out. NULL when tickerSymbol is NULL.	YUM	100%
company_ticker_history	primaryListingExchange	string, nullable	Exchange primaryListingSymbol trades on, canonicalized to one spelling per venue like stockExchange. NULL when tickerSymbol is NULL, and also NULL whenever the link records no venue.	NYSE	95.0%
company_ticker_history	tickerMatchConfidence	float	Confidence score (0-1) of the company-to-ticker match. Links below 0.85 are not shipped.	0.97	100%
company_ticker_history	tickerMatchMethod	string	How the link was established: curated (a reviewed link), website (the employer's own domain matched the issuer's), name_exact, registrant_name (the issuer's registered filer name), name_variant, name_fuzzy, exact_alias (a reviewed alias that links only the one name it lists), corporate_family (inherited from a corporate parent or sibling link), curated_franchise (a reviewed franchise link) or brand_operator (a franchisee or partner whose own name pairs a listed brand with an operator, such as "DWMJ Inc d/b/a McDonald's"). Not every method occurs in every release; the highest-precision subset is curated, website and name_exact.	name_exact	100%
company_ticker_history	validFrom	date, nullable	Start of the listing window. NULL when no listing start is known (windowStartReason unknown or undetermined). The window is closed on both ends.	1996-06-12	51.1%
company_ticker_history	validTo	date, nullable	End of the listing window. NULL means the window is open (the listing is current).	2025-07-16	18.8%
company_ticker_history	listingValidTo	date, nullable	Statutory end of the listing itself (the delisting date), kept beside validTo when a verified acquisition closes the window earlier than the listing formally ended.	2025-07-27	8.9%
company_ticker_history	windowStartReason	string	Why the window opens: first_listing (the listing started on this symbol), symbol_listing (the first listing of the symbol, lent to a link no reviewer approved: weaker evidence than first_listing), first_price_before_corpus (no listing evidence dates the window, so it opens on the security's first price, before January 2022: it means already trading before the panel starts, not listed on that day), symbol_change (the ticker changed and nothing else did), acquisition_close (a verified transaction handed the company to its acquirer), parent_listing (the window starts when the listed parent floated), venue_transfer (the listing moved between two US exchanges: the first trading day on the new one), home_listing_change (the issuer's home listing began, ended or moved; the US line is unchanged), unknown (no listing start was found) or undetermined (the window could not be placed and its start is undated). Empty on a few windows with no recorded reason.	first_listing	99.6%
company_ticker_history	windowEndReason	string	Why the window ends: current (still open), delisted, delisted_to_otc (the listing ended and the same shares kept trading over the counter), acquired (closed on the verified ownership boundary), acquired_date_unknown (an acquirer is known but its start date is not), divested (the issuer sold or spun off this company and kept trading; for a fund holding, the last day the funds held it), symbol_delisted (closed on the delisting date the symbol carries, lent to a link no reviewer approved: weaker evidence than delisted), symbol_change, venue_transfer, home_listing_change or undetermined.	acquired	100%
company_ticker_history	tickerCategory	string	What kind of business the listed company is: staffing, job_board, marketplace, franchisor or operating (every other company). Keep tickerCategory = 'operating' to leave out staffing firms, job boards, marketplaces and franchisors, as a cross-sectional study usually does. The same value ships on every ticker file.	operating	100%
company_ticker_history	tickerTenancyID	string	Opaque 16-character identifier of the listed security this listing window belongs to: one value per security's stay on a symbol. The ticker files key their series on it, so a symbol that passed from one company to another is two series. Empty when the window is tied to no single security.	5480551e25efb958	100%
company_ticker_history	issuerCik	string	The SEC filer (CIK) behind the listing window, where the window is tied to one SEC registrant. NULL otherwise, including foreign listings.	1041061	89.9%

Google Maps Business Data: 77 fields, 203,045,739 rows (catalog 8.3, measured 6 Oct 2026)

Module	Field Name	Type	Description	Example	Coverage
places_listings	cid	string	Google Maps place identifier. Primary key.	9825404506245605297	100%
places_listings	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.	Tops Friendly Markets	11.9%
places_listings	address	string	Full address as the listing shows it.	98 South Ave, Hilton, NY 14468, United States	98.8%
places_listings	amenities	array<string>	Amenities and capabilities present at the place (for example delivery, takeout, wheelchairEntrance), as the listing reports them; only those reported as available appear.	['wheelchairEntrance', 'wheelchairParking', 'wheelchairRestroom']	49.5%
places_listings	avgReviewRating	float	Average review rating recomputed from the review set (alongside the listed reviewRating).	3.88	21.9%
places_listings	businessName	string	The place's name as listed on Google Maps.	Tops Friendly Markets	100%
places_listings	categories	array<string>	Full set of Google Maps categories assigned to the place.	['Supermarket', 'Grocery store']	99.3%
places_listings	category	string	Primary Google Maps business category (e.g. restaurant).	Supermarket	99.3%
places_listings	cbsaCode	string	US metro area code (CBSA), the same metro code the job postings carry. NULL outside the US, and where a US place's location maps to no metro area.	40380	9.3%
places_listings	cbsaTitle	string	Official OMB name of the CBSA in cbsaCode, e.g. New York-Newark-Jersey City, NY-NJ. NULL exactly when cbsaCode is NULL.	Rochester, NY	9.3%
places_listings	companyDomain	string	Normalized registrable domain of website (e.g. acme.com). The join key to this table's companyDomain (derived by the same rule from the company website). NULL when website is empty or unparseable.	topsmarkets.com	31.6%
places_listings	companyStatus	string	Ownership / sector classification of the linked company, always populated (public / subsidiary_of_public / franchisee_of_public / partner_of_public / portfolio_company_of_public / government / nonprofit / educational / private / unknown). unknown when the place has no company link.	private	100%
places_listings	finalCity	string	City, parsed from the address.	Hilton	97.3%
places_listings	finalCountry	string	ISO 3166-1 alpha-2 country code, uppercase (e.g. US); backfilled from the place timezone when the address parse left it empty.	US	100%
places_listings	finalState	string	State or region, as the listing writes it.	NY	72.8%
places_listings	geohash5	string	Geohash of the coordinates at precision 5 (about 5 km cells); a city-centroid cell where the place has no usable coordinate.	dr8wj	100%
places_listings	geohash7	string	Geohash at precision 7 (about 150 m cells); empty where the place has no usable coordinate.	dr8wjzg	100%
places_listings	hasDetailedProfile	integer	1 when the place was collected at least once in detailed mode, else 0. Only these places carry the detailed profile (star histogram, popular times, owner and similar fields).	1	100%
places_listings	imageCount	integer	Count of photos attached to the place listing. NULL when the listing's photos were not collected (unknown, not zero).	3	29.7%
places_listings	inferredEntityType	string	The linked company's hiring role (real_employer, staffing_agency, job_board_aggregator, not_a_company or unknown), from Canaria's company classification. unknown when the place has no company link.	real_employer	100%
places_listings	inferredEntityTypeConfidence	float	Per-field confidence (0-1) of inferredEntityType. NULL exactly when inferredEntityType is unknown. The score recorded by the model-based company classification where a company was classified; the ticker-link rule's measured precision (0.95), or the organisation-type rule's (0.987 Nonprofit, 0.955 Educational), or the posting-evidence classifier's own score (0.70 or more), where it was not.	1	5.5%
places_listings	is24Hours	integer	1 when the listing says the place is open 24 hours.	0	100%
places_listings	isClaimedBusiness	integer	1 when the listing is a claimed/active Google Business Profile (ownerName present), else 0.	1	100%
places_listings	isPubliclyTraded	integer	1 when the linked company is the listed entity; 0 when owned by a listed parent; NULL when the place has no listed-company link.		1.0%
places_listings	isThirdPartySited	integer	1 when the place belongs to a branded network whose sites are typically run by someone else (parcel lockers, rental dealers, charging stations, kiosks inside other stores), so the linked company is the brand, not necessarily the operator; 0 when not flagged, which does not prove the company operates the site. NULL when the place has no company link. Keep isThirdPartySited = 0 for employer-level location counts.	0	11.9%
places_listings	isUSListing	integer	1 when the linked company's ticker trades on a US exchange (NYSE, NASDAQ, NYSE AMERICAN or OTHER (the bucket for NYSE Arca, Cboe BZX and IEX)), 0 when it trades outside the US. NULL when there is no listed-company link, the link records no exchange, or the ticker trades over the counter (OTC).		0.8%
places_listings	latestReviewDate	date	Publish date (UTC) of the newest review among the reviews collected for the place. Google returns a sample of at most about 8 reviews per place, so this is a lower bound on the place's most recent review, not its exact date. NULL when no collected review carries a date, or no reviews were collected.	2026-04-25	16.3%
places_listings	latitude	float	Latitude of the place, as the listing gives it.	43.2858508	100%
places_listings	longitude	float	Longitude of the place, as the listing gives it.	-77.7899955	100%
places_listings	naicsCode	string	6-digit NAICS-2022 industry code derived from the Google Maps category.	445110	81.6%
places_listings	naicsSector	string	2-digit NAICS sector rollup of naicsCode.	44-45	81.6%
places_listings	naicsTitle	string	Official NAICS-2022 title for naicsCode.	Supermarkets and grocery stores	96.1%
places_listings	openingHours	map<string, string>	Opening hours by day, Monday to Sunday, as "open-close" in local time (read with timezone). Closed days, and days the listing does not show, are absent. is24Hours and totalWeeklyHours carry the 24-hour flag and the weekly total.	{'fri': '06:00-23:00'}	38.0%
places_listings	ownerName	string	Google Business Profile owner name, when the listing is claimed.	Example Veterinary Group	44.5%
places_listings	peakDay	string	Day of week with the highest Google Popular Times busyness.	Friday	7.7%
places_listings	peakHour	integer	Hour of day (0-23) with the highest Google Popular Times busyness.	17	7.7%
places_listings	phone	string	Phone number as the listing displays it.	+1 555-0142	59.0%
places_listings	phoneE164	string	The same number in E.164 international form (for example +14155550100).	+15550142	58.8%
places_listings	placeKind	string	What kind of place the listing is: business (its category maps to a NAICS code), non_business (a category on the curated non-business list: bus stops, housing, parks, places of worship, natural features and the like; naicsTitle reads Not a business there) or unknown (a category not mapped yet, or none). "Businesses only" is placeKind = 'business'.	business	100%
places_listings	postalCode	string	Postal / ZIP code as listed; US places carry the 5-digit ZIP (a ZIP+4 is cut to its base, the way a posting's finalZipcode is).	14468	83.7%
places_listings	reviewCount	integer	Number of reviews the listing shows.	579	51.8%
places_listings	reviewCountWithText	integer	Number of reviews carrying written text.	8	22.1%
places_listings	reviewRating	float	Star rating the listing shows (1 to 5).	4.1	56.2%
places_listings	reviewSentiment	string	Overall skew of the review distribution: positive / neutral / polarized / negative. NULL when not computed.	positive	20.4%
places_listings	stars1	integer	Count of 1-star ratings in the place's review histogram.	21	20.4%
places_listings	stars2	integer	Count of 2-star ratings in the place's review histogram.	23	20.4%
places_listings	stars3	integer	Count of 3-star ratings in the place's review histogram.	78	20.4%
places_listings	stars4	integer	Count of 4-star ratings in the place's review histogram.	189	20.4%
places_listings	stars5	integer	Count of 5-star ratings in the place's review histogram.	268	20.4%
places_listings	status	string	Normalized place-operational status (operational / closed_temporarily / closed_permanently / unknown), derived from raw Google hours text.	operational	100%
places_listings	stockExchange	string	Exchange the linked ticker trades on, uppercase and canonicalized to one spelling per venue: US (NYSE/NASDAQ) and foreign markets alike. NULL when tickerSymbol is NULL.		0.8%
places_listings	tickerMatchConfidence	float	Confidence (0-1) of the company-to-ticker link. NULL when there is no listed-company link.		1.0%
places_listings	tickerMatchMethod	string	How the link was established: curated (a reviewed link), website (the employer's own domain matched the issuer's), name_exact, registrant_name (the issuer's registered filer name), name_variant, name_fuzzy, exact_alias (a reviewed alias that links only the one name it lists), corporate_family (inherited from a corporate parent or sibling link), curated_franchise (a reviewed franchise link) or brand_operator (a franchisee or partner whose own name pairs a listed brand with an operator, such as "DWMJ Inc d/b/a McDonald's"). Not every method occurs in every release; the highest-precision subset is curated, website and name_exact.		1.0%
places_listings	tickerParentName	string	Verified registered name of the listed entity that owns the ticker. NULL when there is no listed-company link.		0.8%
places_listings	tickerRelation	string	How the company relates to the linked ticker: self (the listed company), former_name (the listed company under an earlier name), subsidiary, division or acquisition (owned by the listed company), franchise (an independent franchisee operating under the listed company's brand, no equity link), partner (a contracted operator under the brand, such as a delivery or freight partner, no equity link), fund_portfolio (a company held by funds a listed manager runs) or bdc_holding (a portfolio investment of a listed business development company). NULL when there is no listed-company link.		1.0%
places_listings	tickerSymbol	string	Stock ticker of the linked company (e.g. AAPL). NULL when the place has no listed-company link.		1.0%
places_listings	timezone	string	Place timezone; also the source used to backfill finalCountry when the address parse leaves it empty.	America/New_York	100%
places_listings	businessNameLanguage	string	Detected language of businessName.	en	99.9%
places_listings	totalRatings	integer	Total number of star ratings, stars1 to stars5 summed.	579	20.4%
places_listings	totalWeeklyHours	float	Scheduled open hours per week, from openingHours.	17	43.0%
places_listings	website	string	Place website (Google redirect wrappers unwrapped).	https://example-hvac.com	31.6%
places_attributes	cid	string	Google Maps place identifier; joins 1:1 to places_listings.cid.	9825404506245605297	100%
places_attributes	googlePlaceID	string	Google Place ID of the listing. Null when the collection lacks it.		32.9%
places_attributes	plusCode	string	Google Plus Code (open location code) of the listing. Null when the collection lacks it.		99.8%
places_attributes	priceLevel	integer	Google price tier on the 1-4 icon scale ($ to $$$$). NULL when Google shows none.		1.5%
places_attributes	priceCurrency	string	Currency of the listed price range, when Google shows one.		1.5%
places_attributes	priceMinimum	float	Lower bound of the listed price range, when Google shows one.		1.3%
places_attributes	priceMaximum	float	Upper bound of the listed price range, when Google shows one.		1.5%
places_attributes	popularTimesByHour	map<string, array<integer>>	Popular Times histogram: day of week to 24 hourly busyness values (0-100 Google scale), hour 0 first. Empty when the listing shows none.		7.7%
places_attributes	peakBusyness	float	Busyness value (0-100 Google scale) at the peak day and hour.		7.7%
places_attributes	avgWeekdayBusyness	float	Average Popular Times busyness across weekday hours.		7.7%
places_attributes	avgWeekendBusyness	float	Average Popular Times busyness across weekend hours.		7.5%
places_attributes	busiestPeriod	string	Named busiest period of the week (for example weekday lunch, weekend evening) from Popular Times.		7.7%
places_attributes	weeklyBusynessScore	float	Aggregate weekly busyness score computed from the Popular Times curve.		7.7%
places_attributes	menuURL	string	Link to the listing's menu, when it shows one (redirector unwrapped).		0.8%
places_attributes	reservationURLs	array<string>	Reservation links shown on the listing. Empty when none.		0.3%
places_attributes	orderOnlineURLs	array<string>	Order-online links shown on the listing. Empty when none.		1.6%
places_attributes	businessDescription	string	The business's own description text on the listing. Null when none.		1.2%

Salary Benchmark Data: 37 fields, 43,787 rows (catalog 8.3, measured 5 Sep 2026)

Module	Field Name	Type	Description	Example	Coverage
salary_benchmarks	socCode	string	SOC 2018 occupation code. Primary join key.	15-1252	-
salary_benchmarks	socTitle	string	Official SOC 2018 occupation title.	Software Developers	-
salary_benchmarks	state	string	US state two-letter code; 'US' on national rows.	CA	-
salary_benchmarks	granularityLevel	string	'state' or 'national'. Filter on it to pick the rollup you want.	state	-
salary_benchmarks	industry	string	The occupation's most common industry nationally.	Information Technology & Services	-
salary_benchmarks	salaryP10	float	10th percentile of the fused annual salary distribution, USD.	82400	-
salary_benchmarks	salaryP25	float	25th percentile, USD per year.	108300	-
salary_benchmarks	salaryP50	float	Primary benchmark: median annual salary, USD, fused from posted pay, employee-reported pay and BLS wage statistics, weighted by each source's precision.	142500	-
salary_benchmarks	salaryP75	float	75th percentile, USD per year.	187200	-
salary_benchmarks	salaryP90	float	90th percentile, USD per year.	246100	-
salary_benchmarks	ciLower	float	Lower bound of the 95% confidence interval on salaryP50. Wider when the sample is thin or sources disagree.	135400	-
salary_benchmarks	ciUpper	float	Upper bound of the 95% confidence interval on salaryP50.	149700	-
salary_benchmarks	totalPostingsCount	integer	Job postings with a salary behind this cell, of every provenance (employer-stated, source-estimated and model-predicted); the three provenance columns below sum to it. Use it as the fusion's evidence weight, not as a count of employers who published pay.	8420	-
salary_benchmarks	employerStatedPostingsCount	integer	The subset of totalPostingsCount whose salary the employer wrote. The employer-evidence count: use it (plus employeeReportedCount) for your own minimum-sample floor.	3572	-
salary_benchmarks	sourceEstimatedPostingsCount	integer	The subset where the job site supplied its own salary estimate.	295	-
salary_benchmarks	modelPredictedPostingsCount	integer	The subset where Canaria's salary model supplied the figure.	4553	-
salary_benchmarks	employeeReportedCount	integer	Employee-reported salary records behind this cell.	1932	-
salary_benchmarks	sampleSize	integer	totalPostingsCount plus employeeReportedCount. BLS is a reference and does not count here. State rows below 30 are not published.	10352	-
salary_benchmarks	sourcesUsed	array<string>	Which of 'posted', 'reported' and 'bls' fused into this row.	['posted','reported','bls']	-
salary_benchmarks	colaMultiplier	float	Cost-of-living index for the state; 1.0 = US average (always 1.0 on national rows).	1.42	-
salary_benchmarks	nationalEquivalentP50	float	salaryP50 divided by colaMultiplier: the state median expressed in national cost-of-living terms.	100352	-
salary_benchmarks	costTier	string	Cost-of-living bucket: 'High' (multiplier 1.2 or more), 'Medium' (0.85 to below 1.2) or 'Low' (below 0.85).	High	-
salary_skill_premium	skill	string	Canonical skill label, lowercased.	python	-
salary_skill_premium	socCode	string	SOC 2018 occupation code.	15-1252	-
salary_skill_premium	state	string	US state two-letter code.	CA	-
salary_skill_premium	premiumUSD	float	Annual salary difference, USD, for postings listing this skill versus those that do not, in the same occupation and state, weighted across seniority levels. Clamped to plus or minus $100K.	12400	-
salary_skill_premium	premiumPct	float	premiumUSD as a percentage of the without-skill median.	8.5	-
salary_skill_premium	postingsWithSkillCount	integer	Unique salaried postings listing the skill (at least 100 per seniority cell).	3210	-
salary_skill_premium	postingsWithoutSkillCount	integer	Unique salaried postings in the same cells without the skill.	8910	-
salary_yoe_curve	socCode	string	SOC 2018 occupation code.	15-1252	-
salary_yoe_curve	state	string	US state two-letter code (US employee reports only).	CA	-
salary_yoe_curve	yoeBucket	string	Years-of-experience bucket, exactly one of '0-1 year', '1-3 years', '4-6 years', '7-9 years', '10-14 years', '15+ years'. The labels do not sort chronologically as strings; order the curve explicitly.	4-6 years	-
salary_yoe_curve	reportedSalaryP50	float	Median employee-reported annual salary, USD, at this experience bucket.		-
salary_yoe_curve	employeeReportedCount	integer	Employee-reported salary records behind this point (at least 3).		-
salary_yoy_trend	socCode	string	SOC 2018 occupation code.	15-1252	-
salary_yoy_trend	state	string	US state two-letter code.	CA	-
salary_yoy_trend	currentYear	integer	Latest complete calendar year in the window.		-
salary_yoy_trend	currentSalaryP50	float	Median posted annual salary, USD, in currentYear.		-
salary_yoy_trend	priorSalaryP50	float	Median posted annual salary, USD, in the year before currentYear.		-
salary_yoy_trend	yoyChangePct	float	Percentage change from priorSalaryP50 to currentSalaryP50. Cells need at least 100 postings in both years; changes outside -50% to +100% are left out, not clamped.		-
salary_company_profiles	normalizedCompanyName	string	Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.		-
salary_company_profiles	industry	string	The company's most common industry across its postings.		-
salary_company_profiles	distinctSocCount	integer	Distinct SOC occupations the company hires for.		-
salary_company_profiles	totalPostingsCount	integer	The company's US postings with a salary of any provenance, in occupation and state cells of at least 5 postings. Only companies with at least 100 employer-stated salaries have a profile.		-
salary_company_profiles	employeeReportedCount	integer	Employee-reported salary records for the company.		-
salary_company_profiles	transparencyPct	float	Share of the company's postings whose text was observed (postings with a resolved occupation, in US company, occupation and state cells of at least 5 postings) that state a salary in the posting text. Structured salary fields, which some job sites fill with their own estimates, model-predicted salaries and job-site estimates do not count.		-
salary_company_profiles	vsMarketRatio	float	The company's posted median pay against the market median for the same occupations, weighted by its posting mix. Above 1.0 means it pays above market.		-

Job Skills Data: 35 fields, 8,886,228 rows (catalog 8.3, measured 25 Aug 2026)

Module	Field Name	Type	Description	Example	Coverage
skills_title_stats	normalizedTitle	string	Canonical job title (the same value as normalizedTitle on the posting files).	registered nurse	-
skills_title_stats	socCode	string	The title's occupation: the most common SOC 2018 code across the title's postings in the window, the same value on every row of the title. Null when no posting of the title carries a code.	29-1141	-
skills_title_stats	socTitle	string	Official SOC 2018 title of socCode.	Registered Nurses	-
skills_title_stats	socCodeShare	float	Share of the title's postings carrying socCode (0 to 1). Filter on it to drop titles whose occupation is mixed.		-
skills_title_stats	skillType	string	Skill family: technical, soft, certification or license. Part of the key; every statistic is computed within one family.	technical	-
skills_title_stats	skill	string	Lowercased canonical skill string.	python	-
skills_title_stats	postingsWithSkill	integer	Unique postings for this title carrying this skill, counted once per posting.		-
skills_title_stats	postingsWithAnySkillForTitle	integer	The title's unique postings carrying any skill of this family: the denominator for skillCoverage.		-
skills_title_stats	skillCoverage	float	postingsWithSkill divided by postingsWithAnySkillForTitle (0 to 1).		-
skills_title_stats	skillLift	float	Coverage relative to the skill's prevalence across all postings. Above 1 means over-represented in this role.		-
skills_title_stats	skillSpecificity	float	Share of all postings carrying the skill that belong to this title (0 to 1).		-
skills_title_stats	essentialScore	float	Composite score: how often the skill appears on this role's postings, shrinkage-adjusted so low-volume titles are not dominated by noise (0 to 1).		-
skills_title_stats	distinctiveScore	expr:1. / (1. + exp(-distinctive)) over distinctive: float32	Composite score: how strongly the skill marks this role against the whole market, bounded 0 to 1; 0.5 means the market rate, above 0.5 over-represented.		-
skills_title_stats	emergingScore	float	Composite score: growth times confidence. Zero means no growth signal, which is a value, not a gap.		-
skills_title_stats	isStatisticallySignificant	integer	1 when the cell passes a multiple-testing-corrected significance gate (q at or below 0.01 with at least 30 postings).	1	-
skills_trends	normalizedTitle	string	Canonical job title (the same value as normalizedTitle on the posting files).	registered nurse	-
skills_trends	skillType	string	Skill family: technical, soft, certification or license. Part of the key; every statistic is computed within one family.	technical	-
skills_trends	skill	string	Lowercased canonical skill string.	python	-
skills_trends	yearMonth	integer	Calendar month, YYYYMM. A title and skill series has a row only in months where the skill appeared, so a series can contain gaps.	202607	-
skills_trends	monthlyPostingsWithSkill	integer	Unique postings that month for this title carrying this skill.		-
skills_trends	monthlySkillCoverage	float	That month's coverage (0 to 1). Prefer it to raw counts for time series, since monthly posting volumes also reflect collection changes.		-
skills_trends	isPartialMonth	integer	1 on the window's edge months: the first month and the current month are incomplete.	0	-
skills_trends	monthOverMonthGrowth	float	Change in coverage versus the preceding month. Null when the preceding month has no row for the series. Clamped to -1 to +9.		-
skills_trends	yearOverYearGrowth	float	Change in coverage versus the same month a year earlier. Null where the series has no row 12 months earlier. Clamped to -1 to +9.		-
skills_state_trends	state	string	US state two-letter code (the 50 states, DC and the territories). Postings with no resolved state are left out of this file and still count nationally.	TX	-
skills_state_trends	skillType	string	Skill family: technical, soft, certification or license. Part of the key; every statistic is computed within one family.	technical	-
skills_state_trends	skill	string	Lowercased canonical skill string.	python	-
skills_state_trends	yearMonth	integer	Calendar month, YYYYMM.	202607	-
skills_state_trends	monthlyPostingsWithSkill	integer	Unique postings in the state and month carrying the skill. Every skill is kept, so counts add up across months and states.		-
skills_state_trends	monthlyPostingsWithAnySkill	integer	Unique postings in the state and month carrying any skill of this family: the denominator for monthlySkillCoverage.		-
skills_state_trends	monthlySkillCoverage	float	monthlyPostingsWithSkill divided by monthlyPostingsWithAnySkill (0 to 1).		-
skills_state_trends	isPartialMonth	integer	1 on the window's edge months, as on skills_trends.	0	-
skills_state_trends	locationQuotient	float	monthlySkillCoverage divided by the skill's national coverage in the same month: 1.0 is the national share, 2.0 twice as concentrated in the state. Small states make small counts; filter on monthlyPostingsWithSkill before reading it.		-
skills_cooccurrence	skillType	string	Skill family: technical, soft, certification or license. Part of the key; every statistic is computed within one family.	technical	-
skills_cooccurrence	skill	string	Lowercased canonical skill string.	python	-
skills_cooccurrence	coOccurrenceRank	integer	1 to 25: the anchor skill's companions, strongest first. Each skill lists its own top 25, so a pair may appear under either skill or both.	1	-
skills_cooccurrence	relatedSkill	string	The companion skill (same family).		-
skills_cooccurrence	coOccurrencePostingsCount	integer	Postings carrying both skills (minimum 10).		-
skills_cooccurrence	coOccurrenceScore	float	Normalized pointwise mutual information, -1 to 1: 1 = always together, 0 = independent, negative = together less often than chance.		-
skills_global_stats	skillType	string	Skill family: technical, soft, certification or license. Part of the key; every statistic is computed within one family.	technical	-
skills_global_stats	skill	string	Lowercased canonical skill string.	python	-
skills_global_stats	skillPostingsCount	integer	Unique postings carrying this skill across the window.		-
skills_global_stats	distinctJobTitleCount	integer	How many distinct titles list the skill.		-
skills_global_stats	skillPrevalence	float	Share of the family's postings-with-any-skill carrying this skill (0 to 1).		-
skills_global_stats	skillTypeTotalPostingsCount	integer	The family's window total: postings carrying a normalized title and at least one skill of this family, the denominator behind skillPrevalence. A posting counts in every family whose skills it carries, so the family totals overlap.		-
skills_entity_summary	normalizedTitle	string	Canonical job title (the same value as normalizedTitle on the posting files).	registered nurse	-
skills_entity_summary	socCode	string	The title's occupation: the most common SOC 2018 code across the title's postings in the window, the same value on every row of the title. Null when no posting of the title carries a code.	29-1141	-
skills_entity_summary	socTitle	string	Official SOC 2018 title of socCode.	Registered Nurses	-
skills_entity_summary	socCodeShare	float	Share of the title's postings carrying socCode (0 to 1). Filter on it to drop titles whose occupation is mixed.		-
skills_entity_summary	skillType	string	Skill family: technical, soft, certification or license. Part of the key; every statistic is computed within one family.	technical	-
skills_entity_summary	totalPostingsCount	integer	The title's unique postings in the window (the same value on every family row of the title).		-
skills_entity_summary	distinctSkillsCount	integer	Distinct skills retained for the title in this family (capped at 50).		-
skills_entity_summary	topSkills	array<string>	The title's top skills in this family, strongest first.		-

Loading schema explorer...

Data Schema Explorer

Browse every delivered column of the Canaria data products, grouped by delivery module, from delivery catalog 8.3. Select a product below to explore its columns, types, and coverage.

For accuracy benchmarks on every enriched field, see our data quality benchmarks. For definitions of terms like SOC and semantic deduplication, see the glossary.

Job Postings Data: 176 fields, 1,065,010,950 rows (catalog 8.3, measured 6 Oct 2026)

Module

Field Name

Type

Description

Example

Coverage

jobs_core

jobID

string (fixed 64-char key)

SHA-256 per-source posting key: the primary key of the posting files and the join key between them. Changes if the employer edits the title; sourceWebsite with sourceJobKey is the title-edit-stable pair.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

100%

jobs_core

sourceJobKey

string, nullable

The source's own posting identifier, preserved exactly as the source published it. Stable when an employer edits the title; combine with sourceWebsite.

sWtlZMamILRhA845R7pY6Uc4FcPWv_gX9nB0Y-EiQdyOATcAQ-TCRQ

100%

jobs_core

sourceWebsite

string

The source the posting was collected from: aggregated job boards (for example indeed, linkedin, simplyhired, jora, postjobfree) and employer career portals (for example myworkdayjobs, greenhouse, lever, icims, smartrecruiters). Combine with sourceJobKey for joins that survive an employer editing the title. Some sources are retired: google (last posting 2024-08), careerbuilder and jobs2careers (2024-02), kariyer (2024-08) and jora (2025-10); their history stays. reference_source_coverage_monthly gives the monthly state (live, ramp, outage, retired) of every source with US postings.

simplyhired

100%

jobs_core

jobDate

datetime (UTC), nullable

Posting date for the role (UTC, millisecond precision): the source's published posting date when it supplies one, else the day we first collected the posting. jobDateSource says which; see Notes for using the data in the Guide before building daily or weekly series on this column.

2026-05-01 23:18:35.000

100%

jobs_core

jobDateSource

string

Where jobDate comes from: source (the source published it) or discovery_fallback (the source published none, so jobDate is the day we first collected the posting and equals firstSeenDate).

discovery_fallback

100%

jobs_core

firstSeenDate

datetime (UTC)

First time Canaria observed this posting (UTC, millisecond precision), except where firstSeenFromBackfill = 1.

2026-05-01 23:18:35.000

100%

jobs_core

firstSeenFromBackfill

integer

1 when firstSeenDate is not our observation but the posting's own date, taken from an archive to fill a 2024 gap in our PostJobFree collection. Such a row was not known to us on that date: leave it out of point-in-time work keyed on firstSeenDate.

100%

jobs_core

lastSeenDate

datetime (UTC)

Most recent time we observed this posting as still active (UTC, millisecond precision): the later of our last visit to the posting's page and the last time a search-results page or employer career portal still listed it. A listing never moves firstSeenDate. A duration computed from it is a lower bound on how long the posting was open.

2026-05-02 22:21:55.000

100%

jobs_core

contentID

string (fixed 64-char key), nullable

SHA-256 content-version hash over title + description. Shifts when either changes.

e296b88342730dfa05e7e7b5c29ca9c77e4d8187d3012639234abe9ad07e76b7

92.7%

jobs_core

isUniqueJobPosting

integer

1 if this posting represents its group of duplicates (the earliest seen), 0 if it duplicates another. SUM(isUniqueJobPosting) gives unique volume; COUNT() gives gross volume. A relisting first seen more than 60 days after the group's latest sighting starts a new unique posting; shorter gaps keep chaining. Postings that could not be tested count as unique. See Gross and unique postings in the Guide.

100%

jobs_core

duplicateGroupID

string (fixed 64-char key), nullable

Near-duplicate group key: postings sharing it are the same role posted on several sources or collected again over time, and the one with isUniqueJobPosting = 1 is the group's representative. NULL when the posting could not be tested for duplicates, because its location was not resolved precisely enough to compare it.

1de0b412ca7d479c32e343b2f80a6aafc6b27f6e2decff6ff8af342421a5aa98

>99.9%

jobs_core

duplicateGroupSize

integer, nullable

Number of postings in this near-duplicate group (repost intensity). 1 for a posting with no duplicates; NULL when de-duplication has not been applied.

>99.9%

jobs_core

representativeJobID

string (fixed 64-char key), nullable

The jobID of the group's representative posting (the earliest-seen instance, the one with isUniqueJobPosting=1). Points into the same key space as jobID, so a duplicate row joins straight back to its canonical row on representativeJobID = jobID. Equal to the row's own jobID for a representative; NULL when de-duplication has not been applied.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

>99.9%

jobs_core

isDedupProvisional

integer

Deduplication disclosure flag: 1 when the posting's near-duplicate block was left unresolved (block exceeded the resolution cap), so its duplicate grouping is provisional. Exclude flagged rows from deduplication-sensitive metrics for strict analyses.

100%

jobs_core

jobTitle

string, nullable

Raw posting title as published by the employer.

Health & Beauty TL

>99.9%

jobs_core

jobURL

string, nullable

Direct URL to the posting on the source site. The customer-clickable link to the job.

https://www.simplyhired.com/job/sWtlZMamILRhA845R7pY6Uc4FcPWv_gX9nB0Y-EiQdyOATcAQ-TCRQ

99.3%

jobs_core

atsApplicationURL

string, nullable

URL to the employer's ATS application form (Greenhouse, Lever, Workday, etc.) when the posting links out to one. May differ from jobURL when the posting lives on an aggregator. NULL when the only link the source gives is back to the job board itself (for example a LinkedIn posting's own linkedin.com/jobs/view/.. page): that is the board, not an application form.

https://careers.smartrecruiters.com/Dominos

2.9%

jobs_core

companyName

string, nullable

Company name as published on the posting, before normalization. companyName never has leading or trailing whitespace (including non-breaking spaces); a name that is empty after trimming, or HTML markup that cannot be stripped cleanly, is NULL.

Walmart

98.3%

jobs_core

companyProfileURL

string, nullable

URL of the employer's profile or company page on the source site (e.g. an Indeed company page, a LinkedIn company URL), when available.

https://www.linkedin.com/company/travel-nurse-source

27.0%

jobs_core

postedJobFunction

string, nullable

The job function the source site listed for the posting, exactly as stated (for example "Information Technology"). The source's own label, not a Canaria classification; it ships beside socCode, never in place of it. Filled mainly on LinkedIn postings and a few career portals; NULL where the source lists none.

Health Care Provider

18.7%

jobs_core

postedIndustry

string, nullable

The industry the source site listed for the posting's employer, exactly as stated (for example "Staffing and Recruiting"). The source's own label for this posting, so it can differ between postings of one company; it ships beside the NAICS columns, never in place of them. Filled mainly on LinkedIn and a few career portals; NULL where the source lists none.

Hospitals and Health Care

20.7%

jobs_core

postedDepartment

string, nullable

The department or team the source listed for the posting, as stated. A department list sent as a JSON array ships as the department names joined by " | ", and a trailing " /" separator is removed; every other value is exactly as stated. Filled mainly on employer career portals; NULL where the source lists none.

Clinical

0.2%

jobs_core

postingExpirationDate

date, nullable

Reserved for a later release; empty on every row of this release.

jobs_core

isUrgentHire

integer, nullable

Indeed postings only: 1 when Indeed marked the posting as urgently hiring, 0 when it did not. Indeed showed the mark on postings collected from 2022-04 to 2023-06 and from 2026-10; NULL elsewhere. New in catalog 8.3: it arrives with the next build.

jobs_core

isHighVolumeHiring

integer, nullable

Indeed postings only: 1 when Indeed marked the posting as high-volume hiring, 0 when it did not. Indeed showed the mark on postings collected from 2022-04 to 2023-06 and from 2026-10; NULL elsewhere. New in catalog 8.3: it arrives with the next build.

jobs_core

postingLanguage

string, nullable

ISO-style language tag detected on the description (en, es, fr, de, etc.).

92.1%

jobs_core

postingChannel

string

Where the posting was sourced: employer_ats (the employer's own applicant-tracking system, e.g. Workday/Greenhouse/Lever, so a direct employer), job_board (aggregator/marketplace such as Indeed/LinkedIn), or unknown.

job_board

100%

jobs_core

postedLocation

string, nullable

Location string exactly as the posting stated it, before parsing into the final columns. Source HTML markup is removed (each tag becomes a space, edges trimmed); NULL where removing the tags would leave a half-parsed string or nothing at all.

Moraine, OH

91.2%

jobs_core

finalCity

string, nullable

Resolved city name in normalized form.

Moraine

98.0%

jobs_core

finalState

string, nullable

Resolved state or top-level administrative region. US postings carry the 2-letter code (CA, WA); non-US postings carry the canonical region name.

97.2%

jobs_core

finalCountry

string, nullable

Resolved ISO 3166-1 country, uppercase (US, GB, DE). US territories (Puerto Rico, Guam, US Virgin Islands, American Samoa, Northern Mariana Islands) are reported as US with the territory in finalState (PR, GU, VI, AS, MP).

99.2%

jobs_core

finalZipcode

string, nullable

Resolved postal code: the posting's own, when the source publishes one at posting level. A posting found by a location search without a ZIP of its own carries the search's ZIP only when that ZIP lies in the metro area (CBSA) of the city the posting names (or, for a city outside any metro, within 25 km of it); a posting that names no city we can place takes the search's ZIP as its only location.

45419

87.7%

jobs_core

finalLatitude

float, nullable

Latitude of the resolved location at zip/city precision. NULL when only a state- or country-level location could be resolved (we do not present a coarse centroid as a precise point).

39.70733642578125

97.8%

jobs_core

finalLongitude

float, nullable

Longitude of the resolved location at zip/city precision. NULL when only a state- or country-level location could be resolved.

-84.2156982421875

97.8%

jobs_core

cbsaCode

string, nullable

5-digit OMB Core-Based Statistical Area (metro/micro area) code for the posting's resolved ZIP, e.g. 35620. Assignment is one dominant CBSA per ZIP, derived by resolving the Census 2020 ZCTA-to-county crosswalk against the OMB July 2023 CBSA delineation. ZCTAs approximate the USPS ZIP inventory rather than reproducing it, so coverage is close to but not identical with the full ZIP list. A US posting with no resolved ZIP but a resolved city and state takes its city's CBSA: the one holding a strict majority of the city's ZIPs (a city split across metros gets none). NULL when the posting has neither a ZIP nor a mapped city, the ZIP lies outside any CBSA, the resolved country is non-US, OR the ZIP is absent from the ZCTA-derived source. A NULL is therefore "no metro mapping available", not a positive assertion that the location is rural.

19430

89.1%

jobs_core

cbsaTitle

string, nullable

Official OMB name of the CBSA in cbsaCode, e.g. New York-Newark-Jersey City, NY-NJ. NULL exactly when cbsaCode is NULL.

Dayton-Kettering-Beavercreek, OH

89.1%

jobs_role

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

100%

jobs_role

normalizedTitle

string, nullable

Canonical job title from Canaria's title taxonomy of about 100,000 roles: variant spellings, seniority prefixes and synonyms resolve to one canonical title.

Health and Beauty Clerk

99.8%

jobs_role

normalizedTitleConfidence

float, nullable

Confidence for normalizedTitle, 0 to 1, calibrated to P(correct).

0.28422087

99.8%

jobs_role

seniority

string, nullable

Seniority band classified from title and description: intern, entry, mid, supervisor, senior, staff_principal, director or executive. The classifier never declines: every posting it scores receives one of the eight bands, so a NULL is a posting no classifier scored.

entry

99.9%

jobs_role

seniorityTrack

string, nullable

Career track: ic (individual contributor), manager (people manager), or mixed. Distinguishes IC from manager at the same band.

99.9%

jobs_role

seniorityConfidence

float, nullable

Confidence for seniority (the band only), 0 to 1. The classifier's own score, not calibrated to P(correct): choose a cut from this column's own distribution. socConfidence, by contrast, is calibrated and can be thresholded as a probability.

0.8919606

99.9%

jobs_role

seniorityDetail

string, nullable

Finance/banking sub-band refinement for senior roles (banking_vp, vp, c_suite, banking_md). NULL when not applicable. Disambiguates what "VP" means in banking vs tech.

0.3%

jobs_role

seniorityDetailConfidence

float, nullable

Classifier score for seniorityDetail, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever seniorityDetail is NULL. seniorityTrack ships without a score of its own.

0.9698889

<0.1%

jobs_role

socCode

string, nullable

US SOC 2018 6-digit occupation code, classified from title and description context.

39-5012

99.9%

jobs_role

socTitle

string, nullable

Short label for the SOC code in socCode.

Hairdressers, Hairstylists, and Cosmetologists

99.9%

jobs_role

socConfidence

float, nullable

Confidence for socCode, isotonic-calibrated to P(correct), so it is meant to be read as a probability and thresholded as one. A declined classification is reported differently: socCode, socTitle and this column are NULL together, so a NULL means no occupation was assigned, never an assignment at zero confidence.

0.1808818

99.9%

jobs_arrangement

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

100%

jobs_arrangement

employmentStatus

string, nullable

Employment type: full_time, part_time, contract, temporary, internship, per_diem, volunteer, apprenticeship, residency_fellowship or other. A posting titled as a paid senior officer (vice president, president, a chief officer) never ships as volunteer: where the classifier says volunteer, the field is NULL. other is the classifier's rare catch-all (under 0.01% of postings, in practice executive roles with no stated arrangement).

contract

99.9%

jobs_arrangement

employmentConfidence

float, nullable

Classifier score for employmentStatus, 0 to 1, not calibrated to P(correct); the classifier never abstains, so choose a cut from this column's own distribution. NULL exactly when employmentStatus is NULL.

0.61192876

99.9%

jobs_arrangement

workerClassification

string, nullable

Staffing-market worker classification: w2, 1099, statutory_employee or c2c; in practice a w2 / 1099 split; statutory_employee (about 0.01% of postings, usually at low confidence) and c2c (a handful of postings) are emitted rarely and are not usable for analysis. NULL where the posting names no arrangement: the classifier answers unknown rather than default a contract posting to 1099. For postings that accept several arrangements, corp-to-corp included, see workerClassificationAccepted. Its own score is workerClassificationConfidence.

99.1%

jobs_arrangement

workerClassificationConfidence

float, nullable

Classifier score for workerClassification, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever workerClassification is NULL.

0.9929842

1.1%

jobs_arrangement

workerClassificationAccepted

array<string>

For staffing postings that accept multiple arrangements ("W2/1099/C2C OK"), the multi-hot set of accepted classifications. Empty array when the posting names a single arrangement or gives no signal.

[]

0.2%

jobs_arrangement

industryEmploymentSubtype

string, nullable

Healthcare staffing subtype, classified from the posting text: travel, locum_tenens, telehealth_contractor, research_staff or staff. NULL when not a healthcare staffing posting. Because it reads the text, it also fires on some non-clinical travel roles (travelling merchandisers, field service technicians): combine it with a healthcare socCode (29- or 31-) for a clinical-only count.

travel

12.3%

jobs_arrangement

industryEmploymentSubtypeConfidence

float, nullable

Classifier score for industryEmploymentSubtype, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever industryEmploymentSubtype is NULL.

0.82015777

0.1%

jobs_arrangement

temporarySubtype

string, nullable

For temporary roles (employmentStatus = 'temporary'), the temporary nature (seasonal, project, leave_coverage, fixed_term, unspecified). NULL otherwise.

seasonal

1.4%

jobs_arrangement

temporarySubtypeConfidence

float, nullable

Classifier score for temporarySubtype, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever temporarySubtype is NULL.

0.8829124

<0.1%

jobs_arrangement

isGovernment

integer, nullable

1 if the employer is a government body, 0 / NULL otherwise. Pairs with requiresSecurityClearance.

99.9%

jobs_arrangement

isGovernmentConfidence

float, nullable

Classifier score for isGovernment, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever isGovernment is NULL.

0.920121

1.1%

jobs_arrangement

hasFractional

integer, nullable

1 if the role is fractional (e.g. fractional CFO / CMO), 0 / NULL otherwise. Deterministic text-pattern extraction (fractional / interim / retainer-based executive phrasing), not a model score.

99.9%

jobs_arrangement

hasDolApprentice

integer, nullable

1 if the posting is a US DOL-registered apprenticeship, 0 / NULL otherwise. Deterministic text-pattern extraction (DOL registration / apprenticeship.gov / 29 CFR phrasing), not a model score.

99.9%

jobs_arrangement

hasAcgmePgy

integer, nullable

1 if the posting is an ACGME residency / PGY-N training position, 0 / NULL otherwise. Deterministic text-pattern extraction (PGY-N / ACGME / residency-program phrasing), not a model score.

99.9%

jobs_arrangement

remoteStatus

string, nullable

Remote-work classification: Remote, Hybrid or On-Site. NULL only on postings no classifier scored.

On-Site

99.9%

jobs_arrangement

remoteConfidence

float, nullable

Classifier score for remoteStatus, 0 to 1, not calibrated to P(correct); the classifier never abstains. Values concentrate very high, so a 0.95 cut on this column filters far less than the round number suggests.

0.99875367

99.9%

jobs_arrangement

hybridDaysPerWeek

float, nullable

In-office days per week for hybrid roles, rounded to a whole day (0–7) so it filters cleanly (e.g. = 3). Populated only when remoteStatus = 'Hybrid'; NULL otherwise.

7.5%

jobs_arrangement

hybridSchedule

string, nullable

For hybrid roles, whether the in-office schedule is fixed, flexible or unspecified (the posting was read and states none). NULL on non-hybrid rows, and on the few hybrid rows not yet scored for a schedule.

fixed

7.5%

jobs_arrangement

hybridScheduleConfidence

float, nullable

Classifier score for hybridSchedule, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever hybridSchedule is NULL.

0.54035014

3.3%

jobs_arrangement

onsitePattern

string, nullable

For on-site roles, the work-location pattern: fixed_location, rotating (locum tenens / travel nursing), or mobile (field trades). NULL otherwise.

fixed_location

86.3%

jobs_arrangement

onsitePatternConfidence

float, nullable

Classifier score for onsitePattern, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever onsitePattern is NULL.

0.9484171

0.8%

jobs_arrangement

geoRestrictionLevel

string, nullable

Geographic hiring restriction granularity: none, country, state_or_region, metro, or timezone. NULL when there is no signal. Applies to any work mode.

none

99.9%

jobs_arrangement

geoRestrictionLevelConfidence

float, nullable

Classifier score for geoRestrictionLevel, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever geoRestrictionLevel is NULL.

0.90548044

4.3%

jobs_arrangement

hasFiveDayMandate

integer, nullable

1 if the posting states an explicit five-days-a-week in-office mandate, 0 or NULL otherwise. A text-pattern extraction (five-days-in-office phrasing), not a model score; a mandate also sets remoteStatus to On-Site, never Hybrid.

99.9%

jobs_arrangement

shiftTypes

array<string>

Shift-type phrases extracted (night shift, weekends, rotating shift, etc.).

[]

26.5%

jobs_arrangement

travelRequirements

array<string>

Travel-requirement phrases extracted (up to 25% travel, frequent client travel).

[]

4.2%

jobs_arrangement

shiftPattern

string, nullable

The posting's shift pattern as one value, mapped from shiftTypes: multiple (more than one fixed shift named), day, evening, night, rotating, split, shift_work (shifts, which not stated), compressed or flexible. NULL where the posting names none; 24/7 and weekend mentions alone are not mapped. Recall is limited: many postings that state a schedule in free prose get no value, so read a NULL as not detected rather than no shift.

flexible

15.6%

jobs_arrangement

travelRequirement

string, nullable

How much the job travels, as one value from travelRequirements and any travel percentage the description states: frequent, occasional, required (travel is part of the job, extent not stated) or none (the posting says no travel). NULL where the posting says nothing about travel.

required

4.6%

jobs_arrangement

startTiming

string, nullable

The canonical start-timing answer, fused from the three start-date extractors with priority parsed date, then signal, then label: specific_date (a literal date was parsed; see parsedStartDates), within_1_week, within_2_weeks, flexible, immediate (ASAP folds in). NULL when no start signal exists.

immediate

15.3%

jobs_arrangement

startTimingSource

string, nullable

Which extractor produced startTiming: parsed_date, signal, or label.

label

15.3%

jobs_arrangement

advertiserType

string

Type of advertiser (Direct Employer / Staffing Agency / staffing sub-types). Resolved strongest signal first: (1) Staffing Agency whenever a curated structured company signal identifies a staffing/recruiting firm (review-inferred inferredEntityType = staffing_agency, or its normalized industry is Staffing and Recruiting / Temporary Help Services, or, for an employer with no entity-type annotation (inferredEntityType = unknown), a staffing name: staffing, executive search, search partners, search group or headhunter in the name, or a recruiting word (recruiting, recruitment, recruiter, personnel, placement, talent, manpower, temps, temporaries, employment agency, employment services, interim, workforce) together with a model industry code of 561311 or 561320 at a naicsConfidence of 0.70 or better) AND the posting text is either silent or says Direct Employer; (2) otherwise the value inferred from posting text, if any; (3) Direct Employer when the text was silent and postingChannel = employer_ats; (4) Direct Employer when none of the above fired and the company-grain signal identifies a real operating employer (inferredEntityType = real_employer at an inferredEntityTypeConfidence of 0.70 or better, the floor below which the classification treats the label as unverified). Undetermined when no rung finds evidence either way: no advertiser phrase in the text, not an employer applicant-tracking-system feed, and an employer that is neither a known staffing company nor a confidently labelled direct employer. It is not defaulted to Direct Employer. Full domain: Direct Employer, Staffing Agency, Search/Recruiting Firm, Temp/Contract Staffing Agency, Recruitment Process Outsourcing, Employer of Record, Professional Employer Organization, Undetermined. Rung (1) is the only one that overrides a value a weaker rung produced, and it overrides exactly one value: a staffing-family text value (Staffing Agency, Temp/Contract Staffing Agency, Search/Recruiting Firm, Professional Employer Organization, Employer of Record, Recruitment Process Outsourcing) is never overridden, and no posting on a non-staffing company is affected. Consequence you can rely on: this column never reads Direct Employer on a posting whose inferredEntityType is staffing_agency. The company-grain rungs read the same signal in both directions, so the staffing share computed from this column tracks the one computed from inferredEntityType.

Direct Employer

100%

jobs_skills_extracted

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

100%

jobs_skills_extracted

technicalSkills

array<string>

Technical skills extracted from the description, as the skills taxonomy's canonical labels (Python (Programming Language), Node.js (Javascript Library), Terraform, React, etc.). An ambiguous name carries a qualifier in parentheses, so match the FULL label: has(technicalSkills, 'Python (Programming Language)'), never 'python'. Generic occupational nouns that are not skills (e.g. Operations, Scheduling, Transportation) are removed. A small set of domain terms (Finance, Accounting, Logistics, Transportation, Marketing, Construction) is retained only on relevant occupational roles, and dropped otherwise.

[]

86.5%

jobs_skills_extracted

softSkills

array<string>

Soft skills extracted, as canonical labels (Communication, Leadership, Attention to Detail).

[]

68.6%

jobs_skills_extracted

certifications

array<string>

Certifications named in the posting, as canonical labels: an abbreviation ships under the full name (CPA as Certified Public Accountant, PMP as Project Management Professional Certification), often with the abbreviation in parentheses (Registered Nurse (RN), Basic Life Support (BLS) Certification).

[]

25.8%

jobs_skills_extracted

licenseRequirements

array<string>

Professional license requirements extracted, as canonical labels (Driver's License, RN License, CDL Class A).

[]

21.2%

jobs_skills_extracted

degreeLevels

array<string>

Degree levels the description requires (High School, Associates, Bachelors, Masters, Doctoral, Certificate, Postbaccalaureate Certificate, Postmasters Certificate). A level is read only from degree wording, never from benefit text such as tuition assistance, and a level the occupation never requires is removed unless the text clearly states it as a credential. minimumDegree holds the lowest.

[]

38.7%

jobs_skills_extracted

minimumDegree

string

The lowest degree level required, collapsed from degreeLevels; both ship in this file. The collapse is lossy, so read degreeLevels when more than a floor matters. Empty string when no degree level is mentioned. Overloaded abbreviations are excluded where nothing disambiguates them: "dba" never emits Doctoral, and "j.d." emits it only in a degree shape ("J.D. required"), never in "J.D. Power".

High School

38.7%

jobs_skills_extracted

experienceLevels

array<string>

Experience-level cues mentioned (entry-level, senior, executive). Senior is dropped when the posting's seniority signal confidently indicates an entry or mid-level role, removing false positives such as "senior care" on non-senior postings.

[]

11.4%

jobs_skills_extracted

experienceRequirements

array<string>

Years-of-experience requirements parsed from the description, emitted as canonical labels (5+ years, 2-4 years, 10 years), not verbatim source spans: ranges normalize to N-M years and lead-ins like "minimum"/"at least" are stripped. Not converted to integers. Age minimums ("18 years of age") are excluded.

[]

18.5%

jobs_skills_extracted

languageRequirements

array<string>

Language-requirement phrases, including English when explicitly required (bilingual Spanish, fluent in Mandarin, fluent in English).

[]

6.1%

jobs_skills_extracted

physicalRequirements

array<string>

Physical-requirement phrases (lift 50 lbs, stand for 8 hours).

[]

16.6%

jobs_skills_extracted

employerTaggedSkills

array<string>

The skills the SOURCE listed for the posting as structured tags (Indeed's attribute tags, and the same field where another feed publishes it), exactly as listed: one label, or each element of the source's list. [] where the source listed none. Kept apart from technicalSkills / softSkills, which are extracted from the description and normalized to the taxonomy, so a buyer can tell what the employer tagged from what the text says. Not normalized: two sources can spell one skill differently.

[]

1.2%

jobs_skills_extracted

employerTaggedCertifications

array<string>

The certifications and licenses the source listed for the posting, as employerTaggedSkills describes. Compare certifications, extracted from the description.

[]

0.6%

jobs_skills_extracted

employerTaggedQualifications

array<string>

The qualifications the source listed for the posting (Indeed's qualification tags), as employerTaggedSkills describes.

[]

1.0%

jobs_salary

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

100%

jobs_salary

salaryString

string, nullable

The salary text the values were parsed from, as the posting showed it. NULL when the posting states no pay or the values are an estimate. A posting whose pay text gave no usable figure keeps the text beside NULL values: test the value columns, not the text, for whether a salary shipped.

$19 - $32 an hour

41.8%

jobs_salary

salarySource

string, nullable

Where the salary came from: description (stated in the posting text), source_field (the job board's structured salary field, as published; a board can fill that field with its own estimate without flagging it, so description is the only tier certainly in the employer's words), source_estimate (the job board's own flagged estimate) or predicted (a Canaria model estimate, where the posting states no pay). NULL when no salary value ships. Most resolved salaries are estimates, so read this column on every row.

scraped

90.5%

jobs_salary

salaryPayPeriod

string, nullable

The pay period the employer stated, which the annualSalaryUSD values were annualized from: hour, year, week, month, day, two_week or half_month, and rarely a per-unit rate such as shift or mile. Set only when salarySource is description or source_field; NULL on predicted and source_estimate rows (those are annual estimates, not a stated period), wherever salarySource is NULL, and on the rare stated salary whose text named no period. salaryPayPeriod = 'hour' selects every hourly-stated posting without parsing salaryString.

hour

38.5%

jobs_salary

annualSalaryMinimumUSD

float, nullable

Parsed annualized salary minimum, converted to USD at jobDate's FX rate. NULL if no lower bound parsed, the currency isn't covered, or the date is unknown.

39520

89.5%

jobs_salary

annualSalaryAverageUSD

float, nullable

USD-converted average.

53040

90.5%

jobs_salary

annualSalaryMaximumUSD

float, nullable

USD-converted maximum.

66560

88.7%

jobs_salary

salaryConfidence

float, nullable

For salarySource = 'predicted': the model's own confidence, 0 to 1, derived from its predictive uncertainty. It orders predictions but is not a success rate, and it is not the width of the salary range. On the other tiers it marks provenance: 1.0 for stated pay (description, source_field), 0.9 for job-board estimates (source_estimate). Select tiers with salarySource, not with this column (0.9 is stored as 0.8999999761581421, so an equality test matches nothing). NULL wherever no salary value ships.

90.5%

jobs_salary

salaryCurrencyCode

string, nullable

ISO 4217 currency code of the posting's original salary (USD, EUR, INR, KRW). Use this for filtering/grouping; the shipped annualSalaryUSD values are already converted to USD. NULL whenever no annualSalaryUSD value ships, including when the posting's currency had no exchange rate for its date.

USD

90.5%

jobs_salary

isPaid

integer, nullable

1 for paid roles, 0 for unpaid roles (unpaid internships, volunteer positions), NULL when there is no signal.

99.3%

jobs_salary

isPaidConfidence

float, nullable

Classifier score for isPaid, 0 to 1, not calibrated to P(correct). Filled only on postings scored since this score was added, so its fill rate differs sharply by year (see Notes for using the data in the Guide): do not filter on it across years. NULL wherever isPaid is NULL. Also NULL where a stated wage overrode the classifier: the score was for the class that did not ship.

0.9266869

1.1%

jobs_salary

benefits

array<string>

Employer-provided benefits (401K, PAID TIME OFF, HEALTH INSURANCE, etc.): those mentioned in the description, followed by any the job board listed for the posting that the description does not mention, under the same labels. A posting with no description carries the labels its job board listed. A board label that is not a benefit in this vocabulary ("Flexible schedule", "Remote", "Employee stock purchase plan") adds nothing here.

[]

54.2%

jobs_salary

equity

array<string>

Equity or stock-based compensation mentioned (ESPP, Equity, Stock Options, and the rare Cash or Share Options). Distinct from cash benefits in benefits.

[]

3.8%

jobs_salary

compensation

array<string>

Direct cash compensation beyond base pay mentioned (Commission, Performance Bonus, Overtime Pay, Shift Differential, Hazard Pay, Signing Bonus, Severance Pay). Wages, distinct from employer-provided benefits in benefits.

[]

8.7%

jobs_company

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

100%

jobs_company

normalizedCompanyName

string, nullable

Canonical employer name from Canaria's company database: the join key to the company and ticker files. Normally resolved from the posting's own companyName; where a posting names no employer, it can be recovered from its companyProfileURL, or carried from its duplicate group when every named member names the same company.

Walmart

98.2%

jobs_company

normalizedIndustryName

string, nullable

Canonical industry of the employer, one spelling across the variants sources use.

Retail

82.3%

jobs_company

naicsSector

string, nullable

NAICS 2022 sector (2-digit; the multi-prefix sectors read 31-33, 44-45 and 48-49), from whichever source naicsSource names.

44-45

98.0%

jobs_company

naicsIndustryGroup

string, nullable

NAICS-2022 industry group (4-digit). The first four digits of naicsCode wherever a 6-digit code resolved. Label it from the reference_naics bundle at naicsLevel = 4.

4552

98.0%

jobs_company

naicsCode

string, nullable

NAICS 2022 6-digit industry code. NULL mostly where the classifier declined to place the employer; a few of those postings still carry a sector and title from the curated industry lookup.

455211

98.0%

jobs_company

naicsTitle

string, nullable

Official NAICS 2022 label of the most specific level that resolved.

Warehouse Clubs and Supercenters

98.0%

jobs_company

naicsSource

string, nullable

Where both sources have a code, the model wins at or above 0.30 confidence, whether or not the two agree on the sector, and the curated lookup wins otherwise. NULL when no NAICS could be resolved.

industryMap

98.0%

jobs_company

naicsConfidence

float, nullable

Model confidence for the NAICS code, 0 to 1, calibrated to P(correct). Set on model rows; NULL on curated-lookup rows (the lookup carries no score) and where no NAICS resolved. Filter on naicsSource first.

0.9879

51.7%

jobs_company

sourceEntityType

string, nullable

Source-reported legal or organizational structure of the employer (Public Company, Privately Held, Nonprofit, Government Agency, Educational, Partnership, Self-Owned, Self-Employed). The source-reported companion to the classifier-inferred inferredEntityType.

Public Company

71.0%

jobs_company

companyRevenueMinimumUSD

float, nullable

Lower bound (inclusive) of the employer's annual revenue range, in USD.

10000000000

38.6%

jobs_company

companyRevenueMaximumUSD

float, nullable

Upper bound (inclusive) of the employer's annual revenue range, in USD; NULL for open-ended top buckets.

5000000000

28.2%

jobs_company

companyHeadquartersCountryISO2

string, nullable

ISO 3166-1 alpha-2 country code (uppercase) for the employer's headquarters country.

78.7%

jobs_company

companyHeadquartersCity

string, nullable

Employer headquarters city in normalized form.

Bentonville

78.3%

jobs_company

companyHeadquartersState

string, nullable

Employer headquarters state or top-level administrative region. US and Canadian employers carry the 2-letter code; others carry the canonical region name.

73.0%

jobs_company

tickerSymbol

string, nullable

Ticker of the listed company the employer resolves to (e.g. AAPL): its own listing or its owning parent's. The link is decided once per employer name and listed company, from a reviewed sample of that name's postings, and applies to its postings inside the company's ticker window. NULL on a posting a marketplace rule identifies as a third party's listing on the company's platform, and on one whose link did not yet hold on its first-seen date (see the Guide on public-company links).

WMT

23.3%

jobs_company

isPubliclyTraded

integer, nullable

1 when the employer itself is the listed company (including under a former name), 0 when the employer is owned by a listed parent (subsidiary, acquisition, or division), or carries the linked issuer's ticker without being owned by it, NULL when no public-company link was resolved.

23.3%

jobs_company

stockExchange

string, nullable

Exchange the linked ticker trades on, one spelling per venue: NYSE, NASDAQ, NYSE AMERICAN, OTHER (NYSE Arca, Cboe BZX and IEX) or OTC (over-the-counter quotations) for US listings. NULL when tickerSymbol is NULL or the link records no venue. A symbol is not unique across venues: join on tickerSymbol and stockExchange.

NASDAQ

21.4%

jobs_company

tickerRelation

string, nullable

How the employer relates to the linked ticker: self (the listed company), former_name (the listed company under an earlier name), subsidiary, division or acquisition (owned by the listed company), franchise (an independent franchisee operating under the listed company's brand, no equity link), partner (a contracted operator under the brand, such as a delivery or freight partner, no equity link), fund_portfolio (a company held by funds a listed manager runs) or bdc_holding (a portfolio investment of a listed business development company).

self

23.3%

jobs_company

tickerParentName

string, nullable

Registered name of the company that owns the linked ticker (the employer itself for self and former_name links). NULL when tickerSymbol is NULL, and where the owning parent could not be named. Use tickerRelation and isPubliclyTraded, not this column, to tell a parent attribution from an own listing.

Walmart

21.5%

jobs_company

tickerMatchConfidence

float, nullable

Confidence score (0 to 1) of the employer-to-ticker link. Links ship at 0.85 and above. NULL when tickerSymbol is NULL.

23.3%

jobs_company

tickerMatchMethod

string, nullable

How the link was established: curated (a reviewed link), website (the employer's own domain matched the issuer's), name_exact, registrant_name (the issuer's registered filer name), name_variant, name_fuzzy, exact_alias (a reviewed alias that links only the one name it lists), corporate_family (inherited from a corporate parent or sibling link), curated_franchise (a reviewed franchise link) or brand_operator (a franchisee or partner whose own name pairs a listed brand with an operator, such as "DWMJ Inc d/b/a McDonald's"). Not every method occurs in every release; the highest-precision subset is curated, website and name_exact.

annotation

23.3%

jobs_company

isUSListing

integer, nullable

1 when the linked ticker trades on a US exchange (NYSE, NASDAQ, NYSE AMERICAN or OTHER (the bucket for NYSE Arca, Cboe BZX and IEX)), 0 when it trades outside the US, NULL when there is no ticker link, the link records no venue, or the venue is the over-the-counter market. Filter isUSListing = 1 for a US equity panel.

21.1%

jobs_company

primaryListingSymbol

string, nullable

The company's home / ordinary-share listing symbol, carried beside tickerSymbol rather than instead of it, and resolved as of this posting's own date like the rest of the ticker block. primaryListingSymbol != tickerSymbol means tickerSymbol is not the home listing: an ADR or another secondary line: a sufficient test, not an exhaustive one, and it must not be read as "everything else is a single listing". The home leg exists only where Canaria's curated listing reference happens to hold both legs of one company, and there is no other source for it: the SEC-derived listing-date reference holds neither leg of these pairs. So use the inequality to confirm a secondary line, never to rule one out. NULL when tickerSymbol is NULL.

WMT

23.3%

jobs_company

primaryListingExchange

string, nullable

NASDAQ

21.4%

jobs_company

publicFloatUSD

integer, nullable

The public float (market value of the common equity held by non-affiliates, in USD) of the SEC registrant behind tickerSymbol, from the latest annual report (10-K) filed on or before the posting's date: no report filed after the posting is ever used. NULL where tickerSymbol is NULL or no float applies. A subsidiary posting under its parent's ticker carries the parent's float. Floats tagged in the wrong unit are screened out; a small residue remains (see Notes for using the data in the Guide).

391703475732

18.9%

jobs_company

publicFloatAsOf

date, nullable

The date publicFloatUSD is measured at, as the report states it (the last business day of the issuer's second fiscal quarter). NULL exactly when publicFloatUSD is NULL.

2025-07-31

18.9%

jobs_company

marketCapTier

string, nullable

Market-cap tier from publicFloatUSD: large (10 billion USD or more), mid (2 to 10 billion), small (300 million to 2 billion) or micro (under 300 million). Point in time like publicFloatUSD, so usable as a backtest universe filter. The company files carry no tier, because a current tier on a historical posting would be lookahead.

large

18.9%

jobs_company

companyStatus

string

Ownership or sector class, always populated: public (the listed company itself), subsidiary_of_public (a business it owns), franchisee_of_public or partner_of_public (a franchisee or contracted operator under its brand, no equity link), portfolio_company_of_public (held by funds a listed manager runs), government, nonprofit, educational, private, or unknown. A verified listing outranks the reported sector. It always agrees with the ticker block on the same row: where the block is NULL, it reads the employer's class without a link.

public

100%

jobs_company

inferredEntityType

string

Hiring role of the company: real_employer, staffing_agency, job_board_aggregator (a job board, not an employer), not_a_company, or unknown. A role ships only at 0.70 confidence or more, so unknown means not classified, not doubtful. Sources, in order of precedence: a reviewed classification, a verified stock-ticker link, the source-reported organisation type (Nonprofit, Educational), and a classifier reading the company's own postings. A separate axis from sourceEntityType (legal form) and companyStatus (ownership).

real_employer

100%

jobs_company

inferredEntityTypeConfidence

float, nullable

Per-field confidence (0-1) of inferredEntityType. NULL exactly when inferredEntityType is unknown. For a classified employer this is the score recorded by the model-based company classification; for an employer resolved by the ticker-link rule it is that rule's measured precision (0.95); for one resolved by the organisation-type rule it is that label's measured precision (0.987 for Nonprofit, 0.955 for Educational); for one resolved by the posting-evidence classifier it is that classifier's own score, 0.70 or more.

97.0%

jobs_company

attributesAsOf

date

The UTC day this posting's company attributes (industry name, NAICS, size and revenue bands, headquarters, entity type) were resolved. They are that day's values for the employer, not values as of the posting's date: profiles exist only from 2024-09 on, so a 2022 posting carries its employer as resolved now. The ticker block, public float and market-cap tier are point in time and are not covered by this date.

2026-10-06

100%

jobs_signals

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

100%

jobs_signals

remoteWorkMentions

array<string>

Remote-work phrases extracted (fully remote, work from home, hybrid 3 days in office, etc.).

[]

12.6%

jobs_signals

parsedRemotePercentages

array<string>

Remote-percent strings parsed from the description (verbatim spans, e.g. 3 days in office, 100% remote).

[]

0.7%

jobs_signals

remotePercentageLabels

array<string>

Qualitative remote tags (fully remote, fully onsite, remote-first). Independent of parsedRemotePercentages, not a per-entry pairing.

[]

0.8%

jobs_signals

parsedStartDates

array<string>

Start-date strings parsed from the description (verbatim spans).

[]

4.4%

jobs_signals

startDateLabels

array<string>

Qualitative start-date tags (immediate, ASAP, flexible). Independent of parsedStartDates, not a per-entry pairing.

[]

11.5%

jobs_signals

startDateSignals

array<string>

Start-timing phrases from the signal extractor (Flexible Start, Start Within 1 Week, Start Within 2 Weeks). Finer-grained vocabulary than startDateLabels; both feed the canonical startTiming fusion.

['Start Within 2 Weeks']

0.1%

jobs_signals

employmentTypeMentions

array<string>

Employment types mentioned in the text (full-time, part-time, contract, internship, etc.). Verbatim mentions; employmentStatus is the single classified value.

[]

59.3%

jobs_signals

urgencySignals

array<string>

Urgency-signal phrases (immediate start, urgent hiring, start ASAP).

[]

7.8%

jobs_signals

companyTypes

array<string>

Company-type cues mentioned in the posting (startup, fortune 500, non-profit).

[]

16.5%

jobs_signals

eeocStatement

array<string>

EEOC / equal-opportunity statement categories detected on the posting (one element per statement; a posting can carry several). Empty array when none.

[]

37.6%

jobs_eligibility

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

100%

jobs_eligibility

offersVisaSponsorship

integer, nullable

1 if the posting says the role offers visa sponsorship, 0 if it says it does not, NULL when it says neither. Read from the posting text: application questions and bare headings are not statements, and where a posting both refuses and appears to offer, the refusal wins.

1.9%

jobs_eligibility

requiresSecurityClearance

integer, nullable

1 if the posting states that the role requires a security clearance, 0 if it states that it does not, NULL when the posting makes no such statement. Read from the posting text only, so a NULL is "the posting does not say", never "no clearance needed".

1.5%

jobs_eligibility

citizenshipRequirements

array<string>

Citizenship-requirement phrases extracted from the posting.

[]

5.8%

jobs_eligibility

screeningRequirements

array<string>

Screening-requirement phrases (background check, drug test, credit check).

[]

10.3%

jobs_eligibility

visaSponsorshipMentions

array<string>

Visa-sponsorship phrases extracted from the posting (mentions of H-1B, sponsorship language, etc.).

[]

1.5%

jobs_eligibility

clearanceLevels

array<string>

Security clearance levels named in the posting, as canonical labels; match on the full label. Keyword-extracted and independent of requiresSecurityClearance, so a level can appear on a posting that states no requirement.

[]

1.5%

jobs_contacts

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

100%

jobs_contacts

extractedEmails

array<string>

Email addresses found in the description text.

['careers@example-vet.com']

33.4%

jobs_contacts

extractedPhones

array<string>

Phone numbers found in the description text.

['+1-555-0142']

41.4%

jobs_contacts

extractedURLs

array<string>

URLs found in the description text (distinct from jobURL, the posting's own link).

[]

49.7%

jobs_descriptions

contentID

string (fixed 64-char key), nullable

SHA-256 content-version hash over title + description. Shifts when either changes.

3cffff7ad7d58aaa3356e53ee724dd9dfbc3999657146d03b9561a611e80356f

100%

jobs_descriptions

jobDescriptionCleaned

string

Cleaned full job description text: boilerplate and markup stripped, whitespace normalized. Heavy text is split into this file so jobs_core stays narrow; join to jobs_core on contentID.

Our growing veterinary practice is seeking an associate veterinarian to join our team...

100%

jobs_clinical

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

f300000caf40825596634d529102a76d61157f94c6ae2fbe2d2b5cc55c7bd153

100%

jobs_clinical

complianceRequirements

array<string>

Clinical / regulatory compliance phrases extracted from the description (e.g. HIPAA, CMS, Joint Commission).

[]

11.1%

jobs_clinical

facilityTypes

array<string>

Facility-type mentions extracted from the description (hospital, outpatient clinic, long-term care, skilled nursing facility).

['Community Hospital', 'Magnet Hospital', 'Level II Trauma Center']

55.3%

jobs_clinical

patientPopulations

array<string>

Patient-population mentions (pediatric, geriatric, oncology, behavioral health).

['Pediatric']

52.0%

jobs_clinical

caseloads

array<string>

Caseload-related phrases extracted from the description (patient volumes, ratios).

[]

11.0%

jobs_clinical

credentialingTimelines

array<string>

Credentialing-timeline phrases extracted from the description (onboarding, license verification windows, start-readiness).

[]

0.4%

jobs_changes

jobID

string (fixed 64-char key)

The posting's key; joins to jobs_core.

jobs_changes

changeType

string

'upsert' (the posting is new or its delivered values changed; its full current row is in the other jobs files) or 'removal' (delete it on jobID). A full delivery lists every live posting with its latest change and no removals.

upsert

jobs_changes

changedAt

datetime (UTC)

When the change was recorded (UTC).

jobs_changes

vintage

string

The release the change came from.

jobs_open_duration

jobID

string

The posting's key; joins to jobs_core.

deec6437263367ef8e9a8840000c69951b6f959acdaea5c1874f579fa0e73921

jobs_open_duration

postingStatus

string

closed, open (not yet unlisted for its threshold: still open, or closed too recently to tell), or unmeasurable.

jobs_open_duration

unmeasurableReason

string, nullable

Set only when unmeasurable: not_list_crawled (no listing page ever showed the posting: sources collected from detail pages only, most of 2022-2023, and Indeed from 2025-04 to 2026-07), crawl_stopped (we stopped collecting its source or board before it could be confirmed gone), insufficient_crawl_history (no collection calendar or threshold for its source in this release) or not_yet_measured (published after this release's measurement).

jobs_open_duration

lastListedDate

date, nullable

The last day a listing collection showed the posting.

jobs_open_duration

postingOpenDays

integer, nullable

Calendar days from firstSeenDate to lastListedDate, both included. For closed this is the open duration; for open, the days open so far (right-censored). NULL when unmeasurable.

jobs_open_duration

postingDaysSeen

integer, nullable

Distinct days any of our collection saw the posting.

jobs_open_duration

sightingCoverage

float, nullable

postingDaysSeen over the collected days of its unit between first and last sighting, capped at 1. Below 0.5 the span rests on few sightings: treat postingOpenDays as low confidence. NULL where it cannot be computed.

jobs_open_duration

unlistedCrawledDays

integer, nullable

Collected days of its unit since lastListedDate, through measuredThroughDate. A stricter reader can require more than closureThresholdDays.

jobs_open_duration

closureThresholdDays

integer, nullable

Collected days a posting must go unlisted before it counts as closed. Set per source from that source's own history (the smallest gap after which fewer than 5% of postings reappear, between 3 and 60 days) and re-measured on every release.

jobs_open_duration

closureConfidence

float, nullable

1 minus the share of the source's postings that were listed again after being unlisted this many collected days; 0 for a posting listed on its unit's last collected day. Set for open rows too.

jobs_open_duration

closureKnowableDate

date, nullable

For closed only: the first date on which the verdict could have been known: the day after the threshold's last collected day for an employer board; for a job board, the day after the 14th day of the source's calendar that follows it, because whether a job-board day counts as collected compares it with the 14 days either side.

jobs_open_duration

isLeftCensored

integer, nullable

1 when the posting may have opened before we first saw it, so postingOpenDays is a lower bound.

jobs_open_duration

leftCensorReason

string, nullable

posted_before_first_seen (a posting date from our historical archive precedes our first observation), before_list_crawling (first seen before its unit had a listing collection), list_crawling_start (first seen in its unit's first 7 collected days, when a newly collected board is already full of older postings).

jobs_open_duration

sourceLastCrawledDate

date, nullable

The last day its unit was collected.

jobs_open_duration

measuredThroughDate

date, nullable

The last collected day this release measured through: every verdict is as of this date.

Company Data: 143 fields, 29,114,505 rows (catalog 8.3, measured 6 Oct 2026)

Module

Field Name

Type

Description

Example

Coverage

company_profiles

normalizedCompanyName

string

Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.

amazon

100%

company_profiles

affiliatedCompanyNames

array<string>

Distinct names of companies affiliated with this one on the source profile (e.g. LinkedIn affiliated pages). Empty array when none.

['AWS','Twitch']

0.4%

company_profiles

attributesAsOf

date

The UTC day this row's company attributes (industry, NAICS, size and revenue bands, headquarters and the rest) were resolved: the release day. They are the release's current values, not values as of any posting date.

2026-09-30

company_profiles

companyCEO

string

Chief executive's name where the source profile lists one.

Jane Doe

0.1%

company_profiles

companyDataSources

array<string>

Sources that contributed to this profile (array form of firmographicSources). Each firmographic field is chosen separately, preferring LinkedIn where both sources have a value, else the most recent observation, so one company's industry and size band can come from different sources.

['indeed','linkedin']

87.3%

company_profiles

companyDescription

string

Short company description.

More than 60 years ago, Sam Walton opened the first Walmart in Rogers, Arkansas. Today,...

73.5%

company_profiles

companyDomain

string

Registrable domain of companyWebsite, the buyer join key to web-domain datasets.

amazon.com

60.0%

company_profiles

companyEmployeeCount

integer

Employee count shown on the source company profile, where it shows one. It is sourced separately from the self-reported size band (companyEmployeeSizeBucket) and usually lower, because not every employee is listed: use the size band to size a company and this column for the observed count.

1540000

67.3%

company_profiles

companyEmployeeSizeBucket

string

Self-reported employee size band: 1, 2-10, 11-50, 51-200, 201-500, 501-1000, 1001-5000, 5001-10000 or 10001+.

1001-5000

73.5%

company_profiles

companyEmployeeSizeMaximum

integer

Upper bound (inclusive) of the size bucket; NULL for open-ended buckets like 10001+.

5000

73.3%

company_profiles

companyEmployeeSizeMinimum

integer

Lower bound (inclusive) of the size bucket as an integer. Useful for numeric range filters.

1001

73.5%

company_profiles

companyFollowerCount

integer

Follower count on the source company profile (e.g. LinkedIn). NULL when not disclosed.

2100000

64.7%

company_profiles

companyFoundedYear

integer

Year the company was founded.

1994

29.8%

company_profiles

companyHeadquartersCity

string

Headquarters city (canonical, normalized form).

Mountain View

67.8%

company_profiles

companyHeadquartersCountryISO2

string

ISO 3166-1 alpha-2 code (uppercase) of the headquarters country. A few profiles carry a retired or user-assigned code as the source gives it (AN, CS, XK).

73.4%

company_profiles

companyHeadquartersState

string

Headquarters state or top-level administrative region. US and Canadian companies carry the 2-letter code (CA, WA, ON); others carry the canonical region name. Resolved by the same location rules as a posting's finalState.

45.3%

company_profiles

companyHeadquartersZipcode

string

Headquarters postal / ZIP code, parsed from the HQ location string. NULL when the source did not carry one.

98109

0.6%

company_profiles

companyLastVerifiedDate

timestamp

Most recent time a source profile behind this row was collected.

2026-09-22 12:00:00

13.3%

company_profiles

companyLinkedInURL

string

The company's LinkedIn page (where several match, the one with the most followers).

10.7%

company_profiles

companyLogoURL

string

URL of the company logo image from the source profile. NULL when not present.

https://…/logo.png

77.4%

company_profiles

companyOfficeLocations

array<string>

"City, State, Country" for each office listed on the company's source profiles; empty when none is listed.

[]

5.8%

company_profiles

companyRating

float

Aggregate company rating on the source profile (source scale). NULL when not rated.

4.1

2.6%

company_profiles

companyRevenueMaximumUSD

float

Upper bound (inclusive) of the revenue bucket in USD; NULL for open-ended buckets like more than $10B (USD).

500000000.0

0.4%

company_profiles

companyRevenueMinimumUSD

float

Lower bound (inclusive) of the revenue bucket in USD.

100000000.0

0.5%

company_profiles

companyRevenueRange

string

Annual revenue band in USD: less than $1M, $1M to $5M, $5M to $25M, $25M to $100M, $100M to $500M, $500M to $1B, $1B to $5B, $5B to $10B or more than $10B (each label ends in "(USD)"). Filled where a source profile states revenue.

$1B to $5B (USD)

0.5%

company_profiles

companyReviewCount

integer

Number of reviews behind companyRating. NULL when not disclosed.

18452

2.6%

company_profiles

companySpecialties

array<string>

Distinct self-declared specialties / focus areas from the source profile. Empty array when none.

['Cloud Computing','E-commerce']

3.4%

company_profiles

companyStatus

string

Ownership and sector class, set on every profile: public (the listed issuer), subsidiary_of_public (owned by a listed parent), franchisee_of_public, partner_of_public or portfolio_company_of_public (operating under, contracted to, or held by funds of a listed company that does not own it), government, nonprofit, educational, private, or unknown. A verified listing outranks the reported sector; a company identified as listed or listed-owned whose ticker could not be linked reads public or subsidiary_of_public with no tickerSymbol.

private

100%

company_profiles

companyWebsite

string

Primary company domain.

amazon.com

60.0%

company_profiles

firmographicSourceCount

integer

Number of source profile records that contributed to this row; more records mean better-corroborated firmographics.

100%

company_profiles

firmographicSources

string

Comma-separated source systems that contributed firmographics.

linkedin,indeed

87.3%

company_profiles

firstPostingDate

timestamp

Earliest jobDate of any posting attributed to this company (UTC). Where a source publishes no posting date, jobDate is the day we first collected the posting, so an early firstPostingDate can reflect when collection of that source began. A small number of companies carry a first posting before January 2022, from postings collected before the panel starts.

2022-01-14 00:00:00.000

29.9%

company_profiles

placesAvgRating

float

Average star rating (0-5) across matched Google Maps locations, weighted by each location's review count. NULL under the same rule as placesLocationCount, or when none of the matched locations carry a rating.

4.2

13.6%

company_profiles

placesLocationCount

integer

Number of Google Maps business locations matched to this company. Locations flagged third-party-sited are excluded from every places roll-up. NULL when no location matched or when every matched location was flagged.

1842

17.8%

company_profiles

placesNaicsCode

string

Most common NAICS-2022 industry code across the company's matched Google Maps locations (the modal value). NULL under the same rule as placesLocationCount.

722511

17.0%

company_profiles

placesReviewCount

integer

Sum of customer review counts across all matched Google Maps locations. NULL under the same rule as placesLocationCount.

983204

17.8%

company_profiles

placesTopCategories

array<string>

Most common Google Maps business categories across the company's matched locations, most frequent first. Empty when no location matched or every matched location was flagged.

['Coffee shop','Cafe']

17.7%

company_profiles

placesTotalWeeklyHours

float

Sum of weekly open hours across all matched Google Maps locations, a coarse proxy for physical-footprint scale. NULL under the same rule as placesLocationCount.

128640.0

13.3%

company_profiles

inferredEntityType

string

real_employer

100%

company_profiles

inferredEntityTypeConfidence

float

Confidence (0-1) of inferredEntityType; NULL exactly when inferredEntityType is unknown. For a classified company it is the classifier's score, 0.70 or more; for a company resolved by the ticker-link rule, that rule's measured precision (0.95); for one resolved by the organisation-type rule, that label's measured precision (0.987 for Nonprofit, 0.955 for Educational).

0.94

20.0%

company_profiles

isPubliclyTraded

integer

1 when this company is the listed entity (including listings under a former name); 0 when it is owned by a listed parent (subsidiary, acquisition, division) or operates under one's brand; NULL when there is no link to a public listing.

0.2%

company_profiles

isUSListing

integer

1 when the linked ticker trades on a US exchange (NYSE, NASDAQ, NYSE AMERICAN or OTHER (the bucket for NYSE Arca, Cboe BZX and IEX)), 0 when it trades outside the US (for example XETRA, LSE, Euronext Paris, TSX). NULL when there is no ticker link, the link records no exchange, or the ticker trades over the counter (OTC). Filter isUSListing = 1 for a US equity panel.

0.1%

company_profiles

lastPostingDate

timestamp

Latest jobDate. Millisecond precision in UTC.

2026-04-01 00:00:00.000

29.9%

company_profiles

lifetimeDistinctSOCCount

integer

Distinct US Standard Occupational Classification codes across the company's lifetime.

100%

company_profiles

lifetimeDistinctStatesCount

integer

Distinct US states the company has hired in.

100%

company_profiles

lifetimePostingsCount

integer

Gross count of all job postings attributed to this company (every posting, before near-duplicate collapse). Approximate, within about 0.5%.

247830

100%

company_profiles

lifetimePostingsWithClearanceCount

integer

Lifetime postings whose description mentions a security clearance requirement.

380

100%

company_profiles

lifetimePostingsWithRemoteCount

integer

Lifetime postings flagged as remote.

38940

100%

company_profiles

lifetimePostingsWithSalaryCount

integer

Lifetime postings whose text or the source site's structured salary field states pay (the same rule as postingsWithAnyPostedSalaryCount in Hiring History). Structured fields can include a site's own estimate; for pay stated in the posting text, use Hiring History's postingsWithSalaryCount. Approximate, within about 0.5%.

142109

100%

company_profiles

lifetimePostingsWithVisaSponsorshipCount

integer

Lifetime postings whose description mentions visa sponsorship, whether offering or declining it.

5021

100%

company_profiles

lifetimeUniquePostingsCount

integer

De-duplicated lifetime posting count: unique postings only (isUniqueJobPosting = 1), so the same role posted on several sources or collected again over time counts once. Approximate, within about 0.5%.

198042

100%

company_profiles

nAffiliatedPages

integer

Count of distinct affiliated companies (see affiliatedCompanyNames).

100%

company_profiles

nCompanyLocations

integer

Count of distinct office locations listed on the source profile.

100%

company_profiles

nCompanyUpdates

integer

Count of distinct company posts / updates captured from the source profile.

340

100%

company_profiles

nEmployeesListed

integer

Count of distinct employees surfaced across the company's source profiles.

1240

100%

company_profiles

nSimilarPages

integer

Count of distinct similar companies (see similarCompanyNames).

100%

company_profiles

naicsCode

string

NAICS 2022 6-digit industry code. NULL where a company can be placed in a sector but not a 6-digit industry (for example "Manufacturing", sector 31-33).

513210

100%

company_profiles

naicsConfidence

float

Model confidence in naicsCode (0-1), calibrated to the probability the code is correct; set only when naicsSource is model. For companies with few records the score runs low, so the code is right more often than it says: set thresholds by company size rather than one cut for the whole file.

0.71

89.0%

company_profiles

naicsIndustryGroup

string

NAICS-2022 industry group (4-digit), the first four digits of the winning code. Populated when the winning source is specific enough to land at the 4-digit level.

5132

100%

company_profiles

naicsSector

string

NAICS-2022 sector code (2-digit, with canonical range form 31-33/44-45/48-49 for multi-prefix sectors). Comes from whichever NAICS source won for this company: a reviewed public-issuer or large-employer code, the model or the curated lookup. NULL when no source places the company.

100%

company_profiles

naicsSource

string

Which side of the resolution above produced naicsCode / naicsSector / naicsIndustryGroup / naicsTitle: model for the classification model, industry_lookup for the curated lookup. NULL when no NAICS was resolved at all.

model

100%

company_profiles

naicsTitle

string

Short label for the most-specific populated NAICS level.

Software Publishers

100%

company_profiles

normalizedIndustryName

string

Canonical industry label, one spelling across the variants sources use. Filter or join on this.

Information Technology & Services

73.5%

company_profiles

primaryListingExchange

string

Exchange primaryListingSymbol trades on, canonicalized to one spelling per venue exactly like stockExchange. NULL when tickerSymbol is NULL, and ALSO NULL whenever the link itself carries no venue: the two halves of the pair do not fill together, so a non-NULL primaryListingSymbol does not imply a venue beside it. Same contract as the jobs source's column of this name.

SIX SWISS

0.1%

company_profiles

primaryListingSymbol

string

The company's home / ordinary-share listing symbol, carried beside tickerSymbol rather than instead of it. primaryListingSymbol != tickerSymbol means tickerSymbol is not the home listing, an ADR or another secondary line, a sufficient test, not an exhaustive one (see the dual-listing note below for the measured coverage). Note isUSListing describes tickerSymbol, not this pair. NULL when tickerSymbol is NULL, matching the jobs source's column of this name.

ZURN

0.2%

company_profiles

relatedCompaniesCount

integer

Number of companies that share substantial job-posting content with this one, published with topRelatedCompanies. 0 when none are known; this release publishes no related companies, so it reads 0 on every row.

100%

company_profiles

similarCompanyNames

array<string>

Distinct names of companies the source lists as similar to this one. Empty array when none.

['Google','Microsoft']

4.8%

company_profiles

sourceEntityType

string

Source-reported legal / organizational structure of the company (normalized). The source-reported companion to the classifier-inferred inferredEntityType.

Public Company, Privately Held, Nonprofit, Government Agency

40.0%

company_profiles

stockExchange

string

Exchange the linked ticker trades on, one spelling per venue (over-the-counter markets read OTC, Xetra reads XETRA, NYSE American reads NYSE AMERICAN). Venues are never merged. NULL when there is no ticker link or the link records no venue.

NASDAQ

0.1%

company_profiles

tickerMatchConfidence

float

Confidence of the company-to-ticker link (0-1). Links below 0.85 are not shipped.

0.97

0.2%

company_profiles

tickerMatchMethod

string

website

0.2%

company_profiles

tickerParentName

string

Verified registered name of the listed entity that owns the ticker. For self / former_name rows this is the company's own registered name; for subsidiaries it is the parent's.

Apple Inc.

0.1%

company_profiles

tickerRelation

string

How the company relates to the linked ticker: self (the listed company), former_name (the listed company under an earlier name), subsidiary, division or acquisition (owned by the listed company), franchise (an independent franchisee operating under the listed company's brand, no equity link), partner (a contracted operator under the brand, such as a delivery or freight partner, no equity link), fund_portfolio (a company held by funds a listed manager runs) or bdc_holding (a portfolio investment of a listed business development company). NULL when there is no listed-company link.

self

0.2%

company_profiles

tickerSymbol

string

Stock ticker of the listed company this company resolves to (the company itself, or its listed parent or acquirer). Set only on the profiles linked to a listed company; NULL otherwise.

AAPL

0.2%

company_profiles

topEmployeeTitles

array<string>

Up to 10 most-frequent employee job titles observed across the company's source profiles (canonical-level topK). Empty array when none.

['Software Engineer','Program Manager']

0.3%

company_profiles

topRelatedCompanies

array<string>

Reserved for a later release; empty on every row of this release.

['audible','twitch']

company_profiles

topRelatedCompaniesAvgSimilarity

array<float>

Reserved for a later release; empty on every row of this release.

[0.91,0.88]

company_profiles

sourceRecordsCount

integer

Number of source records that name this company under any of the spellings that resolve to it, counted across job postings, company-database profile records and employee-review records. Not a posting count: use lifetimePostingsCount / lifetimeUniquePostingsCount for hiring volume.

92341

100%

company_contacts

normalizedCompanyName

string

Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.

banfield pet hospital

100%

company_contacts

contactEmails

array<string>

De-duplicated recruiting and contact email addresses found across the company's postings. The per-posting contacts behind this roll-up ship in Job Postings Data (jobs_contacts).

['careers@example-vet.com','hr@example-vet.com']

61.3%

company_contacts

contactPhones

array<string>

Deduplicated contact phone numbers aggregated across all of the company's postings.

['+1-555-0142']

45.1%

company_contacts

contactURLs

array<string>

De-duplicated application and contact URLs found across the company's postings.

['https://example-vet.com/careers']

30.9%

company_contacts

postingsWithContactsCount

integer

Number of the company's postings that carried at least one extracted contact (support for the aggregation).

100%

company_contacts

contactLastSeenDate

timestamp

Most recent last-seen date across the postings that contributed contacts (freshness signal for the contact set).

2026-06-28 00:00:00.000

100%

company_hiring_activity

companyPostedSalaryTier

string

Coarse tier of the month's average annual salary across its postings, posted or predicted: high (100K USD or more), medium (60K to 100K USD), low (under 60K USD). NULL when no posting of the month has a salary.

high

99.4%

company_hiring_activity

companyStatus

string

Ownership class of the company, always populated: public (the listed company itself), subsidiary_of_public (a business it owns), franchisee_of_public or partner_of_public (a franchisee or contracted operator under its brand, no equity link), portfolio_company_of_public (held by funds a listed manager runs), government, nonprofit, educational, private, or unknown. On a row with ticker credit it follows the credit.

public

100%

company_hiring_activity

distinctSOCCount

integer

Distinct US SOC codes appearing in this month.

100%

company_hiring_activity

distinctStatesCount

integer

Distinct US states the company posted roles in.

100%

company_hiring_activity

growth90DaysVs365Days

float

Ratio of the last 3 months' unique postings to the trailing 12 months' average quarter, both counted only over the sources live in every month of the 12-month window (like-for-like), so the panel's first 11 months carry no value. Near 1.0 is steady hiring, above 1.5 acceleration, below 0.7 a slowdown. NULL when the trailing 12-month total is below 20 unique postings, the company has fewer than 6 months of observed activity, or any month in the window is partial.

1.68

37.4%

company_hiring_activity

inferredEntityType

string

Hiring role of the company, always populated: real_employer, staffing_agency, job_board_aggregator, not_a_company or unknown. A role ships only at 0.70 confidence or more, so unknown means not classified.

real_employer

100%

company_hiring_activity

isLowVolume

integer

1 if the company had fewer than 10 postings this month; its shares (pay stated, remote) rest on few postings.

100%

company_hiring_activity

isPartialMonth

integer

1 if the month is cut short by the data boundary, so its counts are not comparable to a whole month: the panel's first month when collection starts after the 1st, and the month holding the data cutoff. Leave these rows out of a trend; the counts are real but incomplete. The growth columns are already NULL wherever a comparison touches one.

100%

company_hiring_activity

medianDaysOpen

float

Reserved for a later release; empty on every row of this release.

21.5

company_hiring_activity

medianDaysOpenPostingsCount

integer

Reserved for a later release; empty on every row of this release.

company_hiring_activity

monthOverMonthGrowth

float

Like-for-like month-over-month growth of unique postings: this month's unique postings from the sources live in both months, over the previous month's from the same sources, minus 1. A source starting or stopping therefore does not move it. NULL when that prior count is under 10 unique postings, this month has fewer than 5, or either month is partial.

0.142

21.6%

company_hiring_activity

normalizedCompanyName

string

Canonical company name from Canaria's company database: the join key to the company, posting and ticker files.

Walmart

100%

company_hiring_activity

postingsWithAnyPostedSalaryCount

integer

Postings whose text or the source site's structured salary field states pay (salarySource description or source_field, value > 0), among the postings whose text was observed. Includes structured pay that may be a site's own unflagged estimate (see postingsWithSalaryCount), so compare it across time only within one source (company_hiring_activity_by_source).

312

100%

company_hiring_activity

postingsWithClearanceCount

integer

Postings whose description mentions a security clearance requirement, among postings whose text was observed (share over postingsWithContentCount).

100%

company_hiring_activity

postingsWithContentCount

integer

Postings in this row whose text (description) was observed: the denominator of every share above. A posting whose text was never collected cannot show pay, remote work, visa or clearance language, so it is in totalPostingsCount and in no share; dividing by totalPostingsCount would read a gap in description collection as a change in hiring.

455

100%

company_hiring_activity

postingsWithRemoteCount

integer

Flagged as remote work, among the postings whose text was observed (share over postingsWithContentCount).

100%

company_hiring_activity

postingsWithSalaryCount

integer

Postings whose text states the pay (salarySource = 'description' and a resolved annualSalaryAverageUSD above 0), among postings whose text was observed. A source site's structured salary field is not counted here, because some sites fill it with their own unflagged estimate; it is in postingsWithAnyPostedSalaryCount. Predicted salaries and flagged source estimates are in neither. Share = postingsWithSalaryCount / postingsWithContentCount.

241

100%

company_hiring_activity

postingsWithVisaSponsorshipCount

integer

Postings whose description mentions visa sponsorship, whether offering or declining it, among postings whose text was observed (share over postingsWithContentCount).

100%

company_hiring_activity

snapshotMonth

integer

The New York calendar month of the postings' jobDate (the stated posting date; else the first collection day, less any relative "N days ago" the posting shows; else the first collection day), YYYYMM.

202603

100%

company_hiring_activity

topSOCCounts

array<integer>

Parallel posting counts for topSOCs.

[152, 78]

100%

company_hiring_activity

topSOCs

array<string>

Top 10 occupation codes ranked by posting count.

['15-1252', '15-1132']

100%

company_hiring_activity

topStateCounts

array<integer>

Parallel posting counts for topStates.

[201, 89]

99.7%

company_hiring_activity

topStates

array<string>