US traffic camera availability, month by month

Once a month we record how many public traffic cameras each state transportation department lists, how many of them we were able to check, and how many returned no usable image. The measurement is taken the first time the site runs in a new month, so every month is measured at the same point.

A camera counts as not answering when it returns a placeholder notice instead of a road, or an image identical to another camera’s — two different cameras cannot be looking at exactly the same pixels. Agencies keep publishing these as working, which is why the number exists at all.

We never compare two months measured with different detection methods. When the detection is improved it finds faults that used to slip through, and the numbers rise without a single new camera having broken. Each measurement records which method produced it, and the comparison above refuses to run across a change.

The cameras we check are picked by a stable hash of their id, not by name. Agencies put the location inside the name — Georgia’s cameras are labelled by county — so taking the first two hundred alphabetically meant looking at five counties out of a hundred and seventeen. The same sample is followed every month, so what moves has actually moved rather than being a different draw.

Nothing is deleted. The point of this page is the series, and it cannot be reconstructed after the fact.