How these neighborhood scores are calculated

Where the numbers come from

How the score is built

Each measure is converted to a percentile rank against every other eligible neighborhood, then weighted by profile. Percentile rather than a raw value, because a rank is robust to outliers and because it answers the question a buyer actually has: not "how many burglaries" but "compared to where else".

What is deliberately excluded

School test scores. Achievement correlates with household income and therefore with race, so ranking neighborhoods by it reproduces that correlation and presents it as a quality judgment. Only school access counts here, which is a count, not a quality claim.

Demographics of residents. Never an input, in any form, including proxies.

Commercial crime. Offence counts are filtered to categories that affect someone because they live somewhere: burglary, auto theft, theft from a vehicle, assault, robbery, arson, homicide. Larceny and public disorder concentrate in retail districts, and drug and alcohol arrests track how hard an area is policed rather than how risky it is to live in.

Non-residential areas. A place below 400 residents per square kilometre is excluded rather than ranked badly. Dividing airport offences by the number of people who sleep at the airport produces a large number and no information.

Limits

Noise is not yet measured; the national transportation noise map is raster data behind an interactive viewer with no machine interface. School quality is not measured, for the reason above. Crime figures are a rolling window, so they move as the window slides.

Related

• Denver neighborhoods, measured

Last updated 2026-08-15.

Sources. Offenses: Denver Police Department, offences reported in the last 365 days. Walkability: EPA National Walkability Index. Population: EPA Smart Location Database. Schools: NCES Common Core of Data via Urban Institute. Commute: OSRM, driving time to Union Station.