Road Angel BlackSpot
Est. Manchester, 2011
Verified dataset — Tuesday 14 January
Indexes every A-road, B-road and motorway junction in England, Wales and Scotland.
A civic dataset of 218,472 UK road junctions — 89,300 of them carrying an active BlackSpot warning. Built since 2011 from STATS19 police collision records and 1.2 million community-submitted hazard reports, then ranked on a transparent 0–100 score.
Refreshed every weekday at 06:00 from overnight STATS19 feeds and community submissions. The full leaderboard of 89,300 active warnings lives on the interactive map.
Every score on the index is a number between 0 and 100, calculated the same way for every one of the 218,472 junctions in the dataset. The formula is published, version-controlled and reviewed annually by an external road-safety panel.
Every reported collision recorded by police forces in Great Britain is ingested into the dataset within fourteen days. Fourteen years of records — going back to 2011 — give every junction a long-run baseline of how often, and how seriously, vehicles come to grief there.
Near-misses, obscured sightlines, phantom traffic-light phases, diesel spillages, an awkwardly-placed bus stop. Motorists file 47,800 reports a year from the cab, the saddle and the dashboard. Each is geocoded to the nearest junction, then queued for cross-check.
Within 48 hours of submission, every report is matched against the most recent STATS19 feeds for the same junction and grid square. Reports that agree with police logs are kept; reports that don't are held for a second reviewer before publication. Across 1.2 million submissions, the overall confirmation rate is 94%.
Each junction receives a fresh 0–100 score every weekday at 06:00. Older scores are retained in the public archive, so a junction that improves — through engineering works, signal changes, or simply a change in traffic flow — can be seen to climb back down. Nothing in the dataset is ever quietly deleted.
The dataset is independently cited, freely available to local authorities, and feeds directly into the strategic road network review.
In the spring of 2011, David Hargreaves' sister was hit by a van turning right across her lane at a Croydon junction that, on paper, had no recorded collisions in the previous three years. She spent six weeks in hospital. The official data said nothing had happened there. David — by then twelve years into a career as a Metropolitan Police traffic analyst — knew the official data was only ever part of the story.
He left the force that autumn and began building what would become the Road Angel BlackSpot index: a junction-by-junction map that combined the police's STATS19 records with the unrecorded near-misses, sightline problems and signal-timing complaints that only the people driving through every day could describe.
Fourteen years later, the dataset holds 218,472 junctions and 1.2 million community-submitted reports. It feeds the in-cabin alerts of 140,000 dashcam-equipped drivers through a free API, and is cited by 31 local councils in their road safety plans. But the working assumption at its heart is the same one David took from his sister's near-miss: that the official record is incomplete, and the rest of the map has to be drawn by the people who drive it.