Area crime risk assessment

CAP Index CRIMECAST supplies modeled crime-risk scores for buyer-named addresses and geographic tiers through standing reports, API/batch, GIS, and flat-file delivery. The documented outputs include overall and crime-type scores, national/local benchmark dimensions, a 0–2000 scale where higher means greater modeled risk and 100 is benchmark average, five risk bands, MAX spatial/population methodologies, past/current/projected reference years, report date, and release vintage. The scores are statistical risk assessments, not official incident counts, causation, premises-security findings, or predictions that crime will occur. Public evidence does not disclose the fixed training window, uncertainty, or separate model-build identifier. Current terms say the product is not intended for insurance-eligibility or credit-eligibility determinations; that use restriction must accompany discovery by insurance Functions.

Precisely CrimeIndex (US) supplies an annual licensed full file of modeled comparative crime indices for Census block groups, including national indices and six-category national-relative bands, within-state indices, ten current crime categories, block-group boundaries, and a PBKEY-to-block-group lookup. Higher index values mean greater modeled likelihood of criminal activity; the current 2026 release is pinned to reference year 2025. It answers comparative area crime risk, not official incident counts, premises security adequacy, or a prediction for one premises. Exact numeric scale endpoints, category cutpoints, source observation window, and row-level model/version fields are not documented and must remain limitations.