Everything published here is computed from official open data by code. Nothing is estimated, and no figure was written by a language model. You are welcome to use it, including commercially, under CC BY 4.0 with attribution.
Get the data on GitHub Read the method Cite it: 10.5281/zenodo.22165476
The data is archived on Zenodo with a permanent DOI, so a citation to it keeps resolving even if this site changes. The DOI above always resolves to the most recent version.
What is in it
| Dataset | What it holds |
|---|---|
| UK stop-and-search ethnic disparity, by force | Each force's search shares by ethnicity against that ethnic group's share of the force area's own resident population, from Census 2021. Includes the forces we cannot compute, and why. |
| UK stop-and-search ethnic disparity, national | The same measure aggregated across the 36 forces with a publishable denominator, with the shares it derives from. |
| UK stop-and-search outcomes, by force | Searches, no-further-action rate, searches involving removal of more than outer clothing, and how much ethnicity data was missing. |
| UK complaint review outcomes, by force | Complaint cases logged, reviews completed, and how often the force's own decision was overturned. |
| UK deaths following police contact | The IOPC's categories kept separate rather than collapsed, plus the full year series. |
| US police killings, by state | People killed by police 2013 to 2026, rate per million, body camera use, and whether any officer was charged and how it ended. |
| US police killings, by department | The 208 departments with enough records for a figure to mean something. |
| Australia complaints and deaths in custody | Complaints per 100,000 people and per 100 sworn staff, and deaths in police custody by Indigenous status, for all eight states and territories. |
How to cite it
For the derived data:
PoliceComplaint.com. (2026). Police accountability data: United Kingdom, United States and Australia [Data set]. Zenodo. https://doi.org/10.5281/zenodo.22165476
If you rely on an underlying figure rather than our calculation, please also cite the original publisher. Each dataset names its sources, and they are listed in full on the about page.
What we will not claim for it
Most of the inputs are already public. The UK government's Ethnicity Facts and Figures service already publishes stop-and-search rates by ethnicity and police force area from the same Census. Our version is more current and puts outcome and ethnicity on the same row, but it is not the only such figure in existence and we are not going to pretend otherwise.
Equally: a disparity ratio is a measured difference in outcomes, not proof that anyone acted unlawfully. A recorded complaint is an allegation, not a finding. The limits of each dataset are set out in the method, and the ones that would most easily mislead you are stated at the top of it rather than buried.
Found something wrong?
Open an issue on the repository or tell us. We check it against the source and correct it where the point is made out, whoever raises it.
Data last refreshed . Every figure on this page is recomputed from the source datasets, not edited by hand. Found something wrong? Tell us.