Crime statistics are one of the most politically charged datasets in American public data. They're also one of the most misused. We show them, but we don't let them masquerade as something they aren't.
What we don't do
- We do not compute a 1–10 safety score. There's no defensible way to weight violent crime against property crime against fear of crime, and any number we put on it would be doing the user's reasoning for them.
- We do not show neighborhood-level crime data. FBI Crime Data Explorer reports by agency jurisdiction. The most granular layer that's reliably reportable is city/county.
- We do not allow any user to use this data for tenant screening, employment, or insurance underwriting. Those use cases are governed by the Fair Credit Reporting Act and require certified consumer reports.
What we do
We show:
- Violent crime rate per 100,000 (raw)
- Property crime rate per 100,000 (raw)
- 5-year trend
- Reporting agency name and reporting year
- "Data unavailable" for jurisdictions that didn't report
We always show the source URL and let you click through to the FBI's Crime Data Explorer to see the underlying data.
For neighborhood-level signal, we lean on proxies that are publicly auditable: Walk Score, ACS vacancy rates, OpenStreetMap streetlight density, and 911 response times where municipalities publish them.