Source collection · Cross-program enforcement
DOJ press-release corpus, 2020–2026
135,405 Department of Justice press releases, dated October 25, 2019 through September 15, 2026, harvested from the DOJ's public API on July 4, 2026 and swept again on September 15, 2026 as two compressed JSON-lines files: the whole corpus the three enforcement tables were built from.
Measured on the preserved copy
Facts below were measured on 2026-09-12 by streaming the preserved files, after checking that each file's SHA-256 equals the manifest's. They describe the exact bytes the derived tables were built from.
date field, measured on 2026-09-12 on files whose SHA-256 matched the manifest.The numbers behind this chart
| Month | Releases |
|---|---|
| 2019-10 | 368 |
| 2019-11 | 2K |
| 2019-12 | 2K |
| 2020-01 | 2K |
| 2020-02 | 2K |
| 2020-03 | 1K |
| 2020-04 | 786 |
| 2020-05 | 1K |
| 2020-06 | 2K |
| 2020-07 | 2K |
| 2020-08 | 2K |
| 2020-09 | 2K |
| 2020-10 | 2K |
| 2020-11 | 1K |
| 2020-12 | 1K |
| 2021-01 | 1K |
| 2021-02 | 1K |
| 2021-03 | 2K |
| 2021-04 | 2K |
| 2021-05 | 2K |
| 2021-06 | 2K |
| 2021-07 | 2K |
| 2021-08 | 2K |
| 2021-09 | 2K |
| 2021-10 | 2K |
| 2021-11 | 2K |
| 2021-12 | 2K |
| 2022-01 | 1K |
| 2022-02 | 2K |
| 2022-03 | 2K |
| 2022-04 | 2K |
| 2022-05 | 2K |
| 2022-06 | 2K |
| 2022-07 | 2K |
| 2022-08 | 2K |
| 2022-09 | 2K |
| 2022-10 | 2K |
| 2022-11 | 2K |
| 2022-12 | 2K |
| 2023-01 | 2K |
| 2023-02 | 2K |
| 2023-03 | 2K |
| 2023-04 | 2K |
| 2023-05 | 2K |
| 2023-06 | 2K |
| 2023-07 | 1K |
| 2023-08 | 2K |
| 2023-09 | 2K |
| 2023-10 | 2K |
| 2023-11 | 2K |
| 2023-12 | 1K |
| 2024-01 | 2K |
| 2024-02 | 2K |
| 2024-03 | 2K |
| 2024-04 | 2K |
| 2024-05 | 2K |
| 2024-06 | 2K |
| 2024-07 | 2K |
| 2024-08 | 2K |
| 2024-09 | 2K |
| 2024-10 | 2K |
| 2024-11 | 2K |
| 2024-12 | 1K |
| 2025-01 | 2K |
| 2025-02 | 2K |
| 2025-03 | 2K |
| 2025-04 | 2K |
| 2025-05 | 2K |
| 2025-06 | 2K |
| 2025-07 | 2K |
| 2025-08 | 2K |
| 2025-09 | 2K |
| 2025-10 | 379 |
| 2025-11 | 1K |
| 2025-12 | 2K |
| 2026-01 | 2K |
| 2026-02 | 2K |
| 2026-03 | 2K |
| 2026-04 | 2K |
| 2026-05 | 1K |
| 2026-06 | 1K |
| 2026-07 | 119 |
component field.The numbers behind this chart
| Component | Releases |
|---|---|
| Federal Bureau of Investigation (FBI) | 11K |
| Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF) | 6K |
| USAO - Massachusetts | 5K |
| USAO - Florida, Middle | 4K |
| Drug Enforcement Administration (DEA) | 4K |
| USAO - Pennsylvania, Western | 3K |
| USAO - South Dakota | 3K |
| USAO - New Jersey | 3K |
| USAO - Texas, Southern | 3K |
| USAO - West Virginia, Southern | 3K |
- 135,405 records across 2 files, one JSON object per line, fields
url, title, date, component, number, body_text.dateis a Unix timestamp, not a string. The harvest was cut at 2020-01-01 and taken on 2026-07-04, but the field is the date the DOJ record carries: it runs 2019-10-25 to 2026-07-03, with 3,653 records dated before 2020-01-01 — the harvest stopped at the first API page that reached past the cutoff and kept the whole page. One record in the file carries 2026-07-30, a date after the harvest itself; the DOJ has since corrected that record to 2026-06-30. body_textis present on 99.9% of records, averaging 2,848 characters, HTML stripped.- 172 issuing components; the largest: Federal Bureau of Investigation (FBI) (11,271), Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF) (6,137), USAO - Massachusetts (5,379), USAO - Florida, Middle (4,463), Drug Enforcement Administration (DEA) (3,620).
What it is
Every press release the DOJ published in the window, not only pandemic cases: each record carries the DOJ's identifier, date, title, URL, issuing component, and the text of the release. It is kept whole so that any parsed field in the enforcement tables can be checked back to the sentence it came from, and so the tables can be rebuilt with different search terms.
Where it comes from
Collected on July 4, 2026 by paging the DOJ's press-release API (justice.gov/api/v1/press_releases.json) newest-first back to January 1, 2020, fifty releases a page, and trimmed to the fields above. The API is public and rate-limited; the same harvest can be repeated, and releases published after 2026-07-04 are not in this copy.
Files
Two gzip-compressed JSON-lines files, one release per line.
| File | Rows | Size | SHA-256 | Preserved copy | |
|---|---|---|---|---|---|
doj_press_releases_2020-2026_slim.jsonl.gz | 131,293 | 116.1 MB | 30347880e1cf558b682bdf4677f45bc6db99a52b037e17bde03839b6f34b5ab4 | verified · 2019-10-25 → 2026-07-03 | Download (116.1 MB) |
doj_press_releases_2026-07-04_2026-09-15_slim.jsonl.gz | 4,112 | 3.9 MB | 9640a2297e3665f08038b365f65ee78eaa4bd3dbde6b6b3082fe0379817d33d1 | verified · 2026-07-06 → 2026-09-15 | Download (3.9 MB) |
Columns
One JSON object per line with the release's identifier, date, title, URL, issuing component, and body text. The schema file lists no columns because the file is not tabular.
Caveats
- A press release is not the filing. The DOJ's account of a case is a summary written for the public; the charging document and judgment control.
- Coverage ends with the 2026-07-04 harvest, though the dates the records carry run 2019-10-25 to 2026-07-03.
- The corpus is an API harvest and is not a substitute for the operative court filing in a specific case.
Department of Justice press-release corpus, records dated 2019-10-25 through 2026-07-03, harvested from the DOJ API on 2026-07-04, as preserved by Pandemic Darlings, release pandemic-darlings-data-v1-2026-07-10, https://pandemicdarlings.com/data/doj-press-releases-2020-2026/