Explore the 520 sites · Toronto's Speed Cameras

Filter the whole programme and watch the findings recompute. Nothing here is pre-baked: the numbers come from the same file you can download.

The kind of evidence this is. These are observational data from a programme that launched into a pandemic-era traffic collapse. We report associations, carefully tested, and state the uncertainty plainly. The camera-free-year readings are preliminary and sealed before the data existed; a confirmatory re-pull is due October 2026.

Choose what you want to see

Every camera site in the programme is here. Tick the boxes and the numbers underneath recompute on the spot, so you can check a claim rather than take it. Counts beside each option show how many sites that choice would return before you pick it.

One thing worth trying: set What to show to crashes, then step through the four ticket-volume quartiles. The medians scatter without a gradient, and the quartile that wrote the most tickets is not the one with the largest reduction. The formal tests agree that there is no dose-response: Spearman ρ = 0.07 (p = 0.10) across all 520 sites, and a Kruskal-Wallis across the four quartiles at p = 0.48. That absence is why we read a camera's presence as mattering rather than the volume of tickets it writes.

A note on why a number here may not match a number in the report. This explorer uses all 520 sites. Several report sections apply a symmetric zero-crash exclusion, dropping the 47 sites with no crashes recorded during enforcement so that cameras and radar signs are compared on equal terms. Section 9.5's quartile medians of −4 to −6 percent come from that excluded sample; on the full 520 the same quartiles run nearer −13, −15, −11 and −7. The statistical conclusion, no gradient, is the same either way.

The full file, with every column, is on the data page.

How to read the columns

Column	What it is
Crashes/mo before	Monthly crash rate at the site before any camera arrived, over a baseline running back to 2014
During	Monthly crash rate while the camera was enforcing
After	Monthly crash rate after the camera was removed
Change during	Percent change from before to during, each site against itself
Injuries before	Count of injury collisions in the baseline period
Tickets/mo	Speeding tickets per month while enforcing
Tickets total	All speeding tickets across every deployment at that site
Deployments	How many separate times a camera was posted there; 96 sites had more than one

A site counts as testable for the departure question when it has a non-zero crash rate in both periods and at least three months of data after the camera left. That is the 383-site sample the report uses, and it is the denominator behind the 62 percent figure.