CrimeMaps Severity Score — Methodology

This page documents how CrimeMaps.ca computes the CrimeMaps Severity Score (CMSS) — our severity-weighted crime rate modelled on Statistics Canada's Crime Severity Index.

Formula

CMSS = ( Σ (incidents_i × weight_i) / population × 100,000 ) × scale

Where incidents_i is the count of incidents in category i, weight_i is the Statistics Canada 2006-base severity weight for that category, population is the 2025 Statistics Canada postcensal estimate (2021 Census boundaries), and scale normalises the result so the covered-city population-weighted average CMSS reads as ~100 (analogous to Statistics Canada's CSI convention, but relative to CrimeMaps' covered cities rather than all of Canada). A city with CMSS 150 has 50% more severity-weighted crime per capita than the covered-city average; CMSS 50 means half.

Weight table (version 1.0)

Derived verbatim from Statistics Canada Table 1 — Examples of weights for the Crime Severity Index (2006 base year). Categories we collapse (assault, sexual assault) use the published blend of level 1/2/3 weights.

How CMSS differs from Statistics Canada CSI

• Update cadence: Statistics Canada publishes CSI annually, with ~12 months of lag. CMSS updates as our scrapers pull new incidents (monthly / rolling 60-day).
• Data sources: CSI uses the UCR2 survey (every Canadian police service). CMSS uses municipal open-data portals — not every police service publishes openly, so coverage varies.
• Granularity: CSI publishes at Canada, province, and Census Metropolitan Area level. CMSS also publishes at neighbourhood level where we have polygons.
• Offence granularity: CSI uses ~100 UCR2 offence codes. CMSS uses 13 CrimeMaps categories (collapsed at ingest time).

Honest limitations

1. Coverage bias. Cities that publish more offence types tend to rank higher on CMSS. Version 2.0 will normalise for coverage.
2. Classification fidelity. Ambiguous raw offence strings map to "other" (weight 10). Every CMSS display shows the "other" share so readers can judge data quality.
3. Population mismatch. Some police services cover an area larger than the CMA; dividing by CMA population can slightly inflate the rate.
4. Weight base year. v1.0 uses 2006-base weights. Statistics Canada's current in-force weights (Index IV, 2021-2026) are ~17% higher. v2.0 will migrate.

Cities with limited data coverage

These cities are excluded from the CMSS rankings because their open-data feeds either report fewer than 1,000 incidents in the latest year or have a rate below 200 incidents per 100,000 residents — both strong signals that not every offence category is being published. Individual city pages still show absolute counts from the open-data feed as-is.

• Amherstburg, ON — 150 incidents (fewer than 1,000 incidents in the scoring window)
• Aylmer, ON — 46 incidents (fewer than 1,000 incidents in the scoring window)
• Belleville, ON — 579 incidents (fewer than 1,000 incidents in the scoring window)
• Brockville, ON — 385 incidents (fewer than 1,000 incidents in the scoring window)
• Caledon, ON — 4 incidents (fewer than 1,000 incidents in the scoring window)
• Calgary, AB — 2,009 incidents (below 200 incidents per 100,000 residents (likely incomplete reporting))
• Chatham-Kent, ON — 588 incidents (fewer than 1,000 incidents in the scoring window)
• Cobourg, ON — 289 incidents (fewer than 1,000 incidents in the scoring window)
• Coquitlam, BC — 141 incidents (fewer than 1,000 incidents in the scoring window)
• Espanola, ON — 145 incidents (fewer than 1,000 incidents in the scoring window)
• Gananoque, ON — 45 incidents (fewer than 1,000 incidents in the scoring window)
• Greater Sudbury, ON — 107 incidents (fewer than 1,000 incidents in the scoring window)
• Hamilton, ON — 839 incidents (fewer than 1,000 incidents in the scoring window)
• Medicine Hat, AB — 522 incidents (fewer than 1,000 incidents in the scoring window)
• Niagara Region, ON — 15 incidents (fewer than 1,000 incidents in the scoring window)
• Perth South, ON — 7 incidents (fewer than 1,000 incidents in the scoring window)
• Saugeen Shores, ON — 155 incidents (fewer than 1,000 incidents in the scoring window)
• Shelburne, ON — 121 incidents (fewer than 1,000 incidents in the scoring window)
• South Simcoe, ON — 396 incidents (fewer than 1,000 incidents in the scoring window)
• St. Marys, ON — 32 incidents (fewer than 1,000 incidents in the scoring window)
• St. Thomas, ON — 341 incidents (fewer than 1,000 incidents in the scoring window)
• Stratford, ON — 258 incidents (fewer than 1,000 incidents in the scoring window)
• Strathroy-Caradoc, ON — 216 incidents (fewer than 1,000 incidents in the scoring window)
• Timmins, ON — 587 incidents (fewer than 1,000 incidents in the scoring window)
• West Nipissing, ON — 144 incidents (fewer than 1,000 incidents in the scoring window)
• Wingham, ON — 39 incidents (fewer than 1,000 incidents in the scoring window)
• Woodstock, ON — 607 incidents (fewer than 1,000 incidents in the scoring window)

Versioning

CMSS is versioned. Any weight change constitutes a new version. Every CMSS output in our aggregates.json includes the version field.

• v1.0 (2026-04-24) — initial release with Statistics Canada 2006-base weights, 13-category taxonomy, 2025 postcensal population estimates (2021 Census boundaries).

Source

• Severity weights: Statistics Canada — Measuring Crime in Canada: Table 1
• CSI reference values: StatCan Table 35-10-0026-01
• Populations: Statistics Canada annual demographic estimates (2025), 2021 Census boundaries (Table 17-10-0155-01)