How the Safety Score Works

The safety score answers one question: how much public-space crime activity does this spot see, compared to the rest of Saint Paul? It is a rank from 0 (the most activity in the city) to 100 (the least), not a prediction or a guarantee.

  1. Every incident within a fixed quarter-mile of the address over the past 90 days is counted, weighted by how strongly that type of crime signals danger in public space (see the table below) and by how close it is — crime on your block counts in full, while crime out toward the quarter-mile edge fades, so a hotspot a few blocks away doesn't govern your score. A recent window keeps the score tied to current conditions, so a block that has quieted down reads that way. The score always uses this same quarter-mile circle, so it describes the place itself — widening the radius you monitor for alerts doesn't change how safe the spot scores.
  2. The weighted count is divided by the area of that circle, giving a density — a severity-weighted measure of activity per square mile, not a raw incident count.
  3. That density is ranked against a quarter-mile circle centered on every quarter-mile point of Saint Paul — like-for-like, so a quiet street and a shopping district are each compared fairly. A score of 82 means this spot has less weighted crime activity than 82% of the city.

Why incidents are weighted

Raw incident counts treat a stolen bike the same as a shooting. Worse, they treat crimes that say little about street safety — domestic incidents, which happen between people who know each other, or narcotics cases, which mostly reflect where police patrol — the same as crimes that define a dangerous block, like robbery or shots fired. The weights below reflect how much an incident says about ambient public-space risk, not how serious the crime is for its victim.

Homicide10Maximum weight
Discharge8Shots fired — public-space danger by definition
Robbery8Street crime by definition
Agg assault6Violent, frequently public
Arson5Serious property destruction
Burglary4Property invasion
Auto theft3Property crime
Simple assault3Lower-level violence
Rape3Severe harm, but predominantly committed by someone known to the victim — a weak signal of public-space danger. Weighting reflects signal, not severity of the crime.
Theft2Largely petty and retail theft
Agg assault dom2Domestic — says little about street safety
Criminal damage1Property damage
Vandalism1Property damage
Graffiti1Property damage
Narcotics1Largely reflects where police look, not danger to residents
Simple assault dom1Domestic — says little about street safety

Neighborhood (district) scores

The neighborhood pages score whole district councils rather than single addresses, so they are built a little differently from the address score above. Each of Saint Paul's 17 districts is ranked against the other 16 — not against the citywide quarter-mile baseline — which keeps the comparison like-for-like between districts of very different size and shape.

  1. A district's incidents over the past 90 days are risk-weighted with the same weights shown above.
  2. That weighted total is normalized two ways: per square mile of the district's land, and per 1,000 residents (resident counts from Minnesota Compass, 2020 Census boundaries). Land area keeps a large, park- or river-heavy district from looking artificially calm; residents keep a dense district from looking dangerous just because more people live there.
  3. On each measure the district is ranked among all 17; the two ranks are averaged and rescaled to 0–100, where 100 is the quietest. "Lower crime activity than 12 of 16 other districts" is simply that blended rank put into words.

Limitations

  • Only crimes reported to police and published by the City of Saint Paul are counted; locations are approximate (the city rounds addresses to the block).
  • The address score is not adjusted for population or foot traffic — a busy commercial block sees more incidents per square mile than a residential one without necessarily being more dangerous per visit. The neighborhood scores add a per-resident measure, but even that understates exposure where daytime and visitor populations dwarf residents, as downtown's does.
  • Neighborhood scores reflect reported incidents, and reporting rates vary between neighborhoods — a lower score can partly reflect more reporting, not just more crime.
  • The 90-day window is short by design, so a small, normally quiet district can swing several points on a single serious incident.
  • The scores compare places within Saint Paul only; they say nothing about how the city compares to anywhere else.

See the data live on the crime map or per neighborhood on the neighborhood pages.