How Toronto Crime Score sources and scores its data
Toronto Crime Score ranks Toronto’s neighbourhoods on a single safety-oriented leaderboard. This guide documents where those numbers come from, the assumptions baked into the ranking, and how often the data refreshes — written by the studio that built it, so the choices can be checked instead of guessed at.
The data source
Every incident count on Toronto Crime Score comes from one place: the Toronto Police Service’s Public Safety Data Portal, specifically its Neighbourhood Crime Rates open-data dataset, which counts reported major crimes — homicide, shooting, assault, robbery, break and enter, auto theft, and theft — by neighbourhood over a full calendar year.
The population figures used to turn counts into rates come from the Statistics Canada 2021 Census, and the whole pipeline is published under the Open Government Licence – Toronto. Nothing on the leaderboard is scraped from news reports or user submissions: if a number is on the site, it traces back to the portal.
The ranking assumptions
The first assumption is that neighbourhood size matters more than raw totals: every count is normalized to a rate per 100,000 residents, using the census denominators. A handful of break-and-enters means something different in a neighbourhood of 5,000 people than in one of 50,000. The trade-off is that the census counts residents, while some neighbourhoods — the entertainment district most visibly — absorb daytime and weekend crowds their resident population never captures, so their rates read higher than resident-only figures would suggest. The product states this caveat on the leaderboard itself rather than burying it.
The second assumption is that crimes are not interchangeable. The score weights categories by severity: homicide 35%, shooting 25%, assault 12%, robbery 8%, break and enter 6%, auto theft 5%, theft over $5,000 4%, theft from vehicle 3%, and bicycle theft 2%. Those weights are a judgment call, and the point of publishing them is that the judgment is open. Each category’s contribution is capped at 100, so no single category — however bad a year a neighbourhood had — can single-handedly own its score.
The last assumption is the scope: scores position neighbourhoods against each other within Toronto. They are not absolute safety measurements, they do not transfer to other cities, and they say nothing about policing quality or response times.
How often the data refreshes
The site follows the portal’s publication cycle rather than inventing its own. The Toronto Police Service publishes the neighbourhood crime rates dataset annually; the current leaderboard is built on calendar-year 2025 rates, from the dataset the portal published on February 2, 2026. The product stamps that dataset date on the page, so the freshness of any number you see is checkable at a glance.
Annual granularity is a deliberate trade: it matches exactly what the source publishes and keeps comparisons apples-to-apples across neighbourhoods, at the cost of not showing a mid-year shift until the service publishes it.
What the score does not claim
Reported crime is not total crime — reporting rates vary by neighbourhood — and a leaderboard is not advice about where to live. Many things that make a street feel safe, from lighting to transit to services, are not in the dataset at all. The score is a map of reported police data, ranked, and nothing more.
You can browse the leaderboard on torontocrimescore.com, see Toronto Crime Score among all studio products on the projects page, or read the same kind of methodology notes for our other calculators: the parameters behind the water calculator and how Drone Flight Calc validates flight conditions.