# Sources

Last built 2026-10-26 (Asia/Karachi). Every row in `warn_all.csv` carries its own `source_url` and `retrieved_at`. Raw downloads are kept in `data/raw/<STATE>/`.

Politeness: custom User-Agent that names the project and a contact address, 1.5 s between requests, one pass per refresh, `robots.txt` checked for each host (no disallow on the paths used). Where a site blocked us (403, firewall challenge, 503) we stopped and recorded it below. We did not work around any block.

## States included

| State | Rows in dataset | Months covered | Format | Source page | Notes |
|---|---:|---|---|---|---|
| Alaska (AK) | 4 | 2024-12 to 2026-07 | HTML table | https://jobs.alaska.gov/RR/WARN_notices.htm | Small table; Alaska posts few notices. |
| Alabama (AL) | 64 | 2024-11 to 2026-10 | HTML table | https://www.madeinalabama.com/warn-list/ | Alabama Commerce WARN list. |
| California (CA) | 3,186 | 2024-10 to 2026-10 | XLSX (current) + two fiscal-year PDFs | https://edd.ca.gov/en/jobs_and_training/Layoff_Services_WARN | Current-year XLSX plus the FY24-25 and FY25-26 PDF reports. PDF rows were parsed from text and the monthly totals checked against the PDF's own summary (1,506 and 1,593 notices, exact match). Long company names wrapped across lines in the PDF are cut at the first line; PDF rows carry no city. |
| Florida (FL) | 586 | 2024-10 to 2026-10 | Paged HTML (2024, 2025, 2026) | https://reactwarn.floridajobs.org/WarnList/Records?year=2026&page=1 | FloridaCommerce WARN list, retrieved page by page. |
| Iowa (IA) | 210 | 2024-10 to 2026-10 | XLSX, one sheet per year | https://workforce.iowa.gov/ | Amendment rows (updates to earlier notices) are excluded to avoid double counting. |
| Illinois (IL) | 263 | 2024-10 to 2026-10 | Public XLSX export | https://www.illinoisworknet.com/LayoffRecovery | Only rows flagged as WARN notices are kept (the export also holds non-WARN layoff reports). The export has no effective date. |
| Indiana (IN) | 66 | 2024-10 to 2026-10 | HTML table | https://www.in.gov/dwd/warn-notices/current-warn-notices/ | Indiana DWD current notices. |
| Maryland (MD) | 270 | 2024-10 to 2026-10 | HTML tables, one page per year | https://labor.maryland.gov/employment/warn.shtml | Rescinded notices are excluded. |
| North Carolina (NC) | 68 | 2026-01 to 2026-10 | HTML table (2026 summary list) | https://www.commerce.nc.gov/data-tools-reports/labor-market-data-tools/workforce-warn-reports | 2026 only. The 2024 and 2025 listing pages could not be read with a plain HTTP client, so earlier months are missing. |
| New Jersey (NJ) | 236 | 2024-10 to 2026-10 | XLSX archive | https://www.nj.gov/labor/employer-services/warn/ | New Jersey publishes the month a notice was posted, not a day. Those rows have month precision and no day. |
| New York (NY) | 341 | 2024-10 to 2026-10 | Tableau Public CSV (2026) + legacy HTML pages (2024, 2025) | https://dol.ny.gov/warn-notices | The scriptable dashboard export covers 2026 only. 2024 and early 2025 come from the legacy pages, which show the posting date and no employee count. April to December 2025 is missing. |
| Ohio (OH) | 66 | 2026-01 to 2026-10 | CSV (2026) | https://dam.assets.ohio.gov/raw/upload/jfs.ohio.gov/2026/2026-warn-notice.csv | 2026 only. The 2024 and 2025 lists are published as PDFs or were not reachable. Rows labelled UPDATE are excluded as amendments. |
| Oregon (OR) | 89 | 2024-10 to 2026-10 | XLSX via the state's download form | https://ccwd.hecc.oregon.gov/Layoff/WARN | Oregon lists one row per site; rows sharing a tracking number are combined. |
| South Dakota (SD) | 10 | 2025-01 to 2026-08 | HTML table | https://dlr.sd.gov/workforce_services/businesses/warn_notices.aspx | Very few notices. |
| Texas (TX) | 442 | 2024-10 to 2026-10 | XLSX, one file per year | https://www.twc.texas.gov/data-reports/warn-notice | Texas Workforce Commission. |
| Utah (UT) | 39 | 2024-10 to 2026-09 | HTML table | https://jobs.utah.gov/employer/business/warnnotices.html | Few notices. |
| Virginia (VA) | 126 | 2024-10 to 2026-10 | HTML table | https://www.virginiaworks.gov/warn-notices | Company and address are split from one text field, so a small share of cities may be off. |
| Washington (WA) | 271 | 2024-10 to 2026-10 | ASP.NET search results, paged | https://fortress.wa.gov/esd/file/warn/Public/SearchWARN.aspx | Washington ESD. |
| Wisconsin (WI) | 122 | 2024-10 to 2026-10 | JavaScript-rendered tables, rendered in headless Chrome | https://dwd.wisconsin.gov/dislocatedworker/warn/ | Update entries are excluded. Employee counts are the 'affected workers' figure the state shows. |

## States not obtained, or only partly obtained

Agency page URLs below are the landing pages. For the blocked states the exact failing request URLs were not logged. All checks were made on 2026-10-26.

| State | Agency page | Reason |
|---|---|---|
| MO (Missouri) | https://jobs.mo.gov/warn | Bot challenge (Incapsula). Not circumvented. |
| NM (New Mexico) | https://www.dws.state.nm.us/ | Request blocked by the site's firewall. |
| CO (Colorado) | https://cdle.colorado.gov/employers/layoff-separations/layoff-warn-list | HTTP 403 for our client. |
| MI (Michigan) | https://www.michigan.gov/leo/bureaus-agencies/wd/warn | HTTP 403 for our client. |
| TN (Tennessee) | https://www.tn.gov/workforce/general-resources/major-publications0/major-publications-redirect/reports.html | HTTP 403 for our client. |
| KS (Kansas) | https://www.dol.ks.gov/ | HTTP 403 for our client. |
| MA (Massachusetts) | https://www.mass.gov/lists/layoffs-and-closures | HTTP 403 for our client. |
| GA (Georgia) | https://www.tcsg.edu/warn-public-view/ | HTTP 503 on every attempt. |
| PA (Pennsylvania) | https://www.pa.gov/agencies/dli/resources/for-employers/warn | List is built by JavaScript and no table was found when rendered. |
| OH 2024-2025 (Ohio) | https://jfs.ohio.gov/job-workforce-services/job-programs-and-services/submit-a-warn-notice | Earlier years are PDF-only or returned 404. |
| NC 2024-2025 (North Carolina) | https://www.commerce.nc.gov/data-tools-reports/labor-market-data-tools/workforce-warn-reports | Listing pages returned no table to a plain HTTP client. |
| CT, DC, DE, HI, ID, KY, MT, NE, OK, RI, SC (various) | - | Pages returned 404, held only a few rows, or held data years out of date. Not attempted further. |
| AZ, AR, LA, MN, MS, NV, NH, ND, VT, WV, WY, ME (various) | - | Not attempted in this version. |

## What is excluded on purpose

- Amendments, updates and rescissions of earlier notices (IA, MD, OH, WI). They restate a notice already counted. Logged in `data/rejected.csv`.
- Rows with no date or no company name, and rows dated in the future. Also logged.
- Notices older than 2024-10-01.

No row is estimated or invented. Where a state does not publish a field, the field is empty.