Method
How we score a location
Every score comes from the same fixed formula. Research supplies published crime rates, we do the arithmetic, and the same numbers always produce the same score. Every report can show its own working.
Two rates drive the score
Property crime carries 60% of the weighted index and violent crime carries 40%. Both are measured as reported crimes per 1,000 residents per year, for the smallest reporting area that covers the address: a census tract, neighborhood, police beat or ZIP code.
Benchmarked against the nation
Each local rate is divided by a long-run national rate to get an index, where 1.0 means exactly the national average:
- Property crime benchmark: 19.0 per 1,000
- Violent crime benchmark: 3.8 per 1,000
From rates to a 0–100 score
score = clamp(round(50 + 32 × log₂(0.6 × property index + 0.4 × violent index) + trend adjustment), 0, 100)
50 is the national average. The logarithmic scale means each doubling of the blended crime rate adds about 32 points, so the tool keeps working whether a neighborhood is far below or far above the benchmark.
The trend adjustment is the year-over-year change in reported crime multiplied by 0.4, capped at ±8 points. A neighborhood whose crime is falling gets a small credit; one where it is rising gets a small penalty.
The four bands
Low
0 – 29
Moderate
30 – 54
Elevated
55 – 74
High
75 – 100
When the data is thin
If a published rate is unavailable for an area, that factor is treated as exactly the national average (index 1.0) instead of being guessed, and the report tells you so. Each report also states the geography the rates came from and a confidence level based on how specific and recent the data is.
Rates are estimates. Public crime data is published with a lag, reporting practices vary between agencies, and a single rate cannot describe every street.