How BanditKin counts Flock and ALPR cameras

BanditKin › Flock camera map › Methodology

OpenStreetMap. Most records are automated license plate reader (ALPR) cameras that volunteer mappers have added to OpenStreetMap, the open map database, with tags for the operator, manufacturer, mount and the direction the camera faces. BanditKin reads those records from the live map behind every page.

Community reports. Anyone can report a camera on the map using the “Report a camera” button. Reports are held for moderation and only appear once approved, and each report is checked against existing records so the same camera is not counted twice.

The EFF Atlas of Surveillance. The Electronic Frontier Foundation’s public records project documents which agencies run plate-reader programs. It is a records source, not a map, so it is used for the “On the record” section of each page and never added to camera counts.

Nobody who operates a camera is obliged to declare it. Every count on this site is therefore a documented minimum: the number of cameras that have been mapped, not the number that exist.

Within 10 miles of downtown. The headline figure on every city page counts mapped cameras within a 10-mile great-circle radius of the city’s centre point. It is the scope most public camera maps and news reports use, so it can be compared directly with them.

The wider metro. The second figure counts every mapped camera inside a box roughly 50 miles across centred on the same point. Neighbouring metros overlap, so metro figures must never be added together.

State pages. State totals count only cameras that fall inside the state’s boundary polygon, so they can be added together. As of 28 September 2026 the 50 states hold 141,667 mapped cameras.

Ranks. “Ranks 38th of 251” compares the city with the 251 metro areas BanditKin tracks, all counted the same way on the same day.

“Flock” versus ALPR. Flock Safety is the most common vendor, and “Flock camera” has become shorthand for any plate reader. Counts include every mapped ALPR camera; the share tagged Flock Safety is stated separately, and cameras with no manufacturer tag are reported as such.

The full dataset is re-pulled about once a week. A page’s stated date changes only when its own figures changed in that pull; a page whose numbers did not move keeps its earlier date, so the date on a page is always the date its numbers were last read. Each refreshed page states the change since the previous pull in plain words.

The embedded map on every page is live and updates continuously, so the count shown inside the map viewport can differ from the figures in the text, which are fixed to the stated date and scope.

The city and state pages are generated from the dataset by BanditKin’s own software: every figure, rank and table is computed, never estimated, and the explanatory sentences are written from the data for each place. BanditKin uses an AI assistant to help run this pipeline and to check the output; a person reviews the results before pages are published. No page contains figures that were not read from the dataset on the stated date.

Mapping lags reality: cameras that were installed or removed recently may not yet be reflected. A mapped camera is not proof that it is switched on, and the absence of a mapped camera is not proof that a road is unwatched. Many records have no operator or manufacturer recorded, and only cameras with a recorded facing direction can be used for camera-aware routing. Counts can fall as well as rise when mappers correct earlier entries.

If a camera is missing, misplaced or gone, report it on the map with the “Report a camera” button, or email privacy@banditkin.com. Corrections flow into the next weekly pull.

BanditKin is a private location-sharing app whose navigation routes around mapped cameras. It is not affiliated with Flock Safety or any camera operator. Start at the national Flock camera map or pick a state or city from the index below.