City score crime pillar: police-service csi only, 102 cities mapped

OpenStats policy PR 64 governs how crime severity enters City Score. The safety pillar uses only each city’s own Statistics Canada police service Crime Severity Index (CSI) from tables 35-10-0063-01 and 35-10-0026-01. Regional proxies, provincial averages, and parent CMA CSI rows never substitute for a hometown slug. When mapping finds a police service for a city, the study edition places that CSI on the ranking. When mapping fails, the city stays unplaced at median standing inside the composite.

What PR 64 requires

PR 64 keeps crime data at the police-service grain StatCan publishes. A Vancouver row uses Vancouver municipal police CSI. A Duncan row uses the Duncan RCMP municipal detachment CSI. OpenStats keeps each slug on its own service row without regional borrowing, using data/crime-police-service-map.json for explicit matches. That rule prevents a low-crime neighbour from lending standing to an unmapped city.

The 2025 study edition matched 102 cities to a police-service CSI row. Two hundred forty-one hometowns in the wider corpus lacked a match and receive median standing (50%) on the safety pillar for composite math. Those cities still appear on the City Score board when other pillars place them.

Pillar direction and rank

On the safety pillar, higher CSI means worse standing. Rank 1 on the crime study table is the highest CSI among placed cities (Duncan at 336.06 in 2025). Lower CSI produces better standing when the composite converts study rank into percentile standing. Readers who see “rank 94” on Ottawa’s crime pillar should read it as favourable placement among 343 cities, with CSI 50.61 on the underlying study row.

Examples under PR 64

Ottawa carries CSI 50.61 on its municipal police service row. Toronto carries CSI 54.99 on its service row. Duncan carries CSI 336.06 on the RCMP detachment row. Each value enters only its own slug. Suburbs without a mapped detachment stay at median standing until OpenStats adds a police service match in the map file.

Composites with fewer than eight pillars

City Score averages every pillar OpenStats can observe for a slug, then fills missing pillars at median standing. A city with six observed pillars and two median fills still receives a composite rank among 343 cities. Calgary in 2025 runs six observed pillars when property tax stays unplaced; crime can still enter when PR 64 finds a service row. An unplaced safety city runs seven observed pillars plus median crime, or fewer when other topics also lack data.

Median fill keeps absent crime data neutral rather than punitive or generous. A city cannot earn top composite standing from crime alone while every other pillar sits weak, because eight equal weights still apply. PR 64 defines the crime input only; equal weighting across eight pillars stays unchanged on the composite.

Where to verify a city

Open the hometown chart for the matched police service name and CSI history. Compare Ottawa, Toronto, and Duncan on the study table to see low, mid, and high CSI placements under the same policy.

Map file integrity and explicit matches

PR 64 keeps crime data at the police-service grain Statistics Canada publishes. OpenStats uses data/crime-police-service-map.json for explicit matches so a Vancouver row uses Vancouver municipal police CSI and a Duncan row uses the Duncan RCMP municipal detachment CSI. Regional proxies, provincial averages, and parent CMA CSI rows never substitute for a hometown slug under this policy. That rule prevents a low-crime neighbour from lending standing to an unmapped city and prevents a high-crime region from punishing a slug without its own service row.

The 2025 study edition matched one hundred two cities to a police-service CSI row. Two hundred forty-one hometowns in the wider corpus lacked a match and receive median standing on the safety pillar for composite math. Those cities still appear on the City Score board when other pillars place them, which keeps coverage broad while crime inputs remain neutral until mapping improves.

Direction and percentile conversion

On the safety pillar, higher CSI means worse standing. Rank 1 on the crime study table is the highest CSI among placed cities, Duncan at 336.06 in 2025. Lower CSI produces better standing when the composite converts study rank into percentile standing. Ottawa at CSI 50.61 and Toronto at CSI 54.99 each enter only their own slug under PR 64. Suburbs without a mapped detachment stay at median standing until OpenStats adds a police service match in the map file.

Composites with median fills

City Score averages every pillar OpenStats can observe for a slug, then fills missing pillars at median standing. Calgary in 2025 runs six observed pillars when property tax stays unplaced, yet crime can still enter when PR 64 finds a service row. Median fill keeps absent crime data neutral rather than punitive or generous. A city cannot earn top composite standing from crime alone while every other pillar sits weak, because eight equal weights still apply. PR 64 defines the crime input only, while equal weighting across eight pillars stays unchanged on the composite.

Verify any city on the hometown chart for matched police service name and CSI history. Compare Ottawa, Toronto, and Duncan on the study table to see low, mid, and high CSI placements under the same policy. Coverage expansion replaces median fill with observed inputs one slug at a time when the map file gains a vetted match.