Crime statistics are among the most politically charged numbers in American public data, and among the most misused. We show them (they matter to a move) but we won't let them masquerade as something cleaner than they are.
What we won't do
No 1–10 safety score. There is no defensible way to weigh violent crime against property crime against the fear of crime, and any single number we stamped on a city would just be doing your reasoning for you, badly. The tidy score is exactly the thing that makes crime data dangerous.
No neighborhood-level data. The FBI Crime Data Explorer reports by agency jurisdiction. The finest layer that's reliably reportable is city or county, so that's where we stop. A heat map down to the block would look authoritative and be mostly noise.
No screening use. Tenant screening, employment, and insurance underwriting are governed by the Fair Credit Reporting Act and require certified consumer reports. Our data is for understanding a place, not for judging a person, and our terms forbid the latter outright.
What we show instead
The raw violent and property crime rates per 100,000, the reporting agency's name, the reporting year, and a five-year trend, with a clear "data unavailable" wherever a jurisdiction simply didn't report, because pretending zero means "safe" is its own kind of lie. Every figure links straight back to the FBI's own Crime Data Explorer so you can check us.
For finer-grained signal, we lean on proxies anyone can audit: Walk Score, Census vacancy rates, OpenStreetMap streetlight density, and municipal 911 response times where cities publish them. None of those is crime; but together they tell you more about how a street feels than a single fabricated score ever could.
This is the same principle behind everything we publish: show the source, show the date, and let you judge. Compare two places and read the crime numbers the way they're meant to be read: as evidence, not a verdict.