Nason/open-policing-database

Stanford Open Policing Project Database

Every standardized police stop released by the Stanford Open Policing Project: 35 state patrols and 53 city or county departments, roughly 2000-2020, in one stops table with the project's standardized columns typed (date, time, place, subject race/sex/age, outcome, search, contraband, vehicle) and each location's raw_* source columns kept as text. locations documents every source file with its observed and stated coverage and the project's data notes; v_stops_by_year gives per-location annual counts. 230,002,705 rows across 2 tables.

Table	Description	Row Count	Column Count
stops	One row per police stop (traffic and pedestrian) from every Stanford Open Policing Project location, 2000s-2020s: the project's standardized columns (date, time, place, subject demographics, outcome, search, contraband, vehicle) typed, plus each location's raw_* source columns as text; location_id/state/city say where a row came from	230,002,617	223
locations	One row per source file: state, city or statewide, release date, rows loaded, observed first/last stop dates, the data readme's stated coverage and data notes	88	14

Query it remotely

INSTALL httpfs; LOAD httpfs;
ATTACH 'https://huggingface.co/datasets/Nason/open-policing-database/resolve/main/open-policing.duckdb' AS policing (READ_ONLY);

SELECT state, subject_race, COUNT(*) AS stops,
       ROUND(100.0 * COUNT(*) FILTER (WHERE search_conducted) / NULLIF(COUNT(*) FILTER (WHERE search_conducted IS NOT NULL), 0), 2) AS pct_searched
FROM policing.stops WHERE date >= '2015-01-01' AND subject_race IN ('white', 'black', 'hispanic')
GROUP BY 1, 2 ORDER BY 1, 2;

Or with the datapond packages: pip install datapond / pak::pak("datapond-db/datapond-r").

Build pipeline, the project's data readme and caveats: https://github.com/ian-nason/open-policing-database Source: Stanford Open Policing Project, https://openpolicing.stanford.edu (ODC-By 1.0; cite Pierson et al. 2020).