Home › Data Store

Data Store

One-time downloads of six public labor datasets — WARN layoffs, H-1B visas, SEC 8-K filings, bankruptcies, unemployment claims and JOLTS — or a monthly refresh. What you get, where it comes from and its limits are stated below before you pay.

Built for research desks, quant teams and analysts. Available in Parquet for direct integration with trading systems.

For funds: Hedge fund annual contract → · Short-seller signals → · Custom research report →

Format:

One-Time Download Monthly Subscription

One-time = full history, snapshot as of purchase. Monthly = a rolling 1-year window refreshed while you subscribe; full history still needs the one-time purchase.

Want a finished report, not raw CSVs? Custom research reports → · Hedge fund? See the annual contract →

Best Value

Complete Data Bundle

All 6 datasets with full history, as a one-time snapshot on the purchase date. Ongoing refreshes are the monthly option (1-year rolling window, per the toggle above).

85K+ WARN 5.1M LCA 809K H-1B 10K+ SEC 1.4K+ Bankruptcy JOLTS + DOL

$1,499

All 6 datasets included

or buy on Gumroad →

WARN Notices

Daily

85,000+ mass layoff and plant closure records. 47 states and DC (AR, WY, NH do not publish). Company, city, county, employees, dates, industry.

Data card

Source: State workforce-agency WARN listings, one scraper per state + DC; each record keeps the source_url it was captured from — per-notice for most recent records, the state listing page for much of the backlog (source_granularity says which).
Period: Notices from 1988 to the purchase date (depth varies by state; most agencies publish a rolling multi-year list).
Schema: id, company_name (as filed), display_name (normalized), state, city, county, employees_affected, notice_date, effective_date, effective_date_source, layoff_type, naics_code, ticker, source_url, first_ingested_at, last_changed_at, and the other columns of the records table listed in the reference, including scrape bookkeeping (scraped_at, scraper_content_hash); only import-time helper columns are excluded.
Missingness: Some agencies publish no headcount, no city or no effective date; those fields are null, and an estimated effective date is marked effective_date_source = estimated. Counts are not imputed.
Coverage gaps: AR, WY, NH: no public listing. Amendments to city, county, effective date or type update the record in place (last_changed_at); a re-filing with a different headcount or company name becomes a new record, because the id is derived from those fields. Not every layoff produces a WARN notice (federal threshold 100+ employees / 50+ affected).
Refresh: Scrape daily at 05:00 UTC; this file is a snapshot at purchase. Monthly option: rolling 1-year window.
Licensing: Public records, normalized by us; use in your own research and products under the terms. Resale or redistribution of the files needs a separate commercial licence. No accuracy guarantee.
Preview without email: sample rows · three records as filed (captured 2026-09-16)

Preview sample data

or buy on Gumroad →

+ stay current for $49/mo
Monthly refresh of new WARN filings. Cancel anytime.

Sources

Public records only: state workforce agencies (WARN), SEC EDGAR (8-K), USCIS and the Department of Labor (H-1B/LCA, claims, JOLTS), federal bankruptcy courts. Every WARN record keeps its agency source URL and the company name exactly as filed next to the normalized name.

Coverage and freshness

WARN: 47 states and DC publish notices (Arkansas, Wyoming and New Hampshire do not); scraped daily at 05:00 UTC. A download is a snapshot as of the purchase date; the monthly option refreshes it. Cadence per dataset is on the methodology page.

Known limits

Values are as published by the agencies — no accuracy guarantee. Headcounts are as filed; where an agency omits an effective date we estimate one and mark the row (effective_date_source). Company names are normalized for matching; the raw name is kept.

Delivery, refunds, license

One-time download link valid 72 hours, also emailed. Digital goods: a purchase is non-refundable once its link has been accessed, per the terms; for a delivery or file problem, contact support. Use is for your own research and products under the terms — reselling or redistributing the files needs a separate commercial license.

Investment research desks: the alternative-data overview describes how these files are used alongside filings.

Inspect rows without giving an email: sample data · three records as filed.

Try Before You Buy

Download a free sample pack — 500 WARN records, 1,000 LCA records, and 100 H-1B records in CSV. See the schema, coverage, and quality before purchasing the full dataset.

No credit card required. The download link is emailed when you submit; delivery time depends on your mail provider.

Enterprise & Cloud Delivery

Need data delivered via AWS Data Exchange, Snowflake, or Databricks? Contact us for enterprise licensing and custom data feeds.

Frequently Asked Questions

What formats are available?

CSV — universal spreadsheet format. JSON — structured data for applications. Parquet — columnar format optimized for analytics (works with Pandas, Spark, BigQuery). JSON-LD — linked data format with schema.org markup for AI/LLM pipelines.

How is the data delivered?

After payment, you receive a one-time download link (valid 72 hours). Single-dataset purchases download as a single file. Multi-dataset purchases (Complete Bundle, H-1B/LCA, JOLTS/DOL) download as a ZIP archive with one file per dataset. A copy of the download link is also emailed to you.

What columns are included?

All public columns for each dataset. WARN: company, state, city, county, employees affected, notice/effective dates, industry, layoff type, and more. LCA: employer, job title, wage, worksite, SOC code, visa class, dates. H-1B: employer, job, wage, state, fiscal year, status. SEC: ticker, CIK, form type, items, key facts. Bankruptcy: debtor, chapter, court, WARN cross-reference. Internal columns (hashes, normalization fields) are excluded.

Do you offer refunds?

Data purchases are digital downloads: per the terms, a one-time purchase is non-refundable once its download link has been accessed. If you have a delivery or file problem, contact support with your order email.

Need ongoing API access instead?

Check out our API plans starting at $0/month (free tier with 25 calls/day). Pro plans include webhooks, bulk exports, and cross-dataset signals like Risk Signal and Talent Pipeline.