What is CrimeScore?

The CrimeScore is a composite crime intensity index. A higher score means more crime relative to other ZIP codes in the dataset. The raw score is normalized into a percentile — so a ZIP in the 99th percentile has more crime than 99% of all US ZIP codes.

Percentile ranges and their labels:

CrimeScore is not a prediction. It reflects reported crime incidents from police departments, aggregated by SpotCrime. Areas with less police reporting may appear safer in the data — this is a known limitation of any crime data product.

CrimeScore does not model future crime, predict individual risk, or account for unreported incidents. It is a structured representation of available incident data.

All crime data comes from SpotCrime.com, which aggregates police incident reports across the US. SpotCrime has been collecting crime data since 2008 and covers hundreds of millions of reported incidents from hundreds of jurisdictions.

Scores are updated monthly. The “snapshot month” shown on each score represents the month the model was run, incorporating all available data through that period. Historical snapshots are retained so you can observe trends over time.

Approximately 41,700 US ZIP codes (ZCTAs) are currently scored. County-level FIPS scores are also available. If a region shows “not found,” it means that area isn't covered yet — not that it's safe.

Each score includes a model_hash — a cryptographic identifier of the exact model version used to generate it. This ensures scores are reproducible and auditable. If the methodology changes, the hash changes, making it clear which version produced a given score.

SpotCrime.com is the leading public crime mapping platform, with data going back to 2008 and covering hundreds of millions of reported incidents. CrimeScore is SpotCrime's structured intelligence layer — built for developers, AI assistants, and data products.