Road Angel BlackSpot
Est. Manchester, 2011
Verified dataset — Tuesday 14 January
Founded in Manchester · 2011
In the spring of 2011, David Hargreaves spent six weeks at King's College Hospital while his younger sister recovered from a multi-vehicle collision at a Croydon junction he had crossed on his own commute a thousand times. The driver responsible was never charged. The junction layout was never changed. The official report ran to four pages and answered nothing about why the same place kept hurting people.
Hargreaves was then a Metropolitan Police traffic analyst with fourteen years of STATS19 collision data on his desk and a private conviction that the absence of a public, postcode-level hazard map was the single largest gap in British road safety. He resigned, took a flat in Manchester near the universities where the GIS talent lived, and began typing the first rows of what would become the Road Angel BlackSpot register. Fourteen years later, the dataset indexes 218,472 mapped junctions, with 89,300 carrying an active warning.
"I never wanted to build a safety product," he says. "I wanted to build a register — the kind of document a journalist, a council engineer, or a worried sister could open and trust."
The Manchester Desk · 14 staff
Six former collision investigators, two chartered civil engineers, four data analysts, a press officer, and a community editor — based in a single open-plan office on Whitworth Street.
Head of Verification
Twelve years with the Met's Collision Investigation Unit before joining in 2014. Signs off every BlackSpot that enters the active register.
Senior Collision Investigator
Twenty-one years on the GMP collision desk. Reviews the most severe community reports and arbitrates the 6% of submissions that don't match the police record.
Chartered Civil Engineer (MICE)
Joined from Highways England in 2020. Translates STATS19 geometry into the layout features that the BlackSpot Score weighs most heavily.
Lead Data Analyst
Built the 48-hour STATS19 cross-check pipeline. An ex-collision investigator who now writes more SQL than collision reports.
Plus ten colleagues across verification, engineering, data, press and community editing — every named role listed on the Methodology page.
We publish a register. We do not promise safety. We do not sell a product that prevents accidents. We publish what drivers, police and engineers together have observed at a junction, and we let the reader weigh that evidence on their own commute. Where the data is thin, we say so. Where a hotspot cannot be confirmed, we say so. Where a BlackSpot is verified, we mark it plainly and we keep the mark until the evidence retires it.
Editorial Standards
Four cross-checks, two human sign-offs, and an explicit list of things this publication will never claim.
§ 01
Every driver submission is automatically compared against the Department for Transport's STATS19 police collision record within 48 hours of receipt. Where the two records agree, the junction's BlackSpot Score is recalculated and any new contributory factors are added. Where they disagree — roughly 6% of submissions — the report is routed to a senior investigator for a manual decision before any score is changed.
§ 02
Each of the 218,472 mapped junctions carries a score between 0 and 100, weighted from collision frequency, severity mix, contributory-factor profile, layout complexity and the volume of corroborating community reports. The score is recalculated nightly. The methodology is public and is now cited by 31 local councils in their road safety plans.
§ 03
Of the 1.2 million community-submitted hazard reports processed since launch, 94% are confirmed against police logs or corroborated by at least three independent driver reports within 30 days. A confirmed BlackSpot is never published silently — it carries a date stamp, a confidence grade and the name of the investigator who signed it off.
§ 04
We do not publish individual driver names or vehicle details from incident reports. We do not claim that following BlackSpot data will prevent an accident — only that it informs a driver. We do not offer medical or legal advice regarding post-accident claims. We do not assert endorsement by the Department for Transport or any police force unless a written partnership is on file and dated.
Trusted With
National Highways
2019 data partnership feeding the M-road safety review
BBC Watchdog
Featured investigation, 2022
The AA Magazine
Long-form citation in the Spring 2023 commuter issue
UK Transport Awards 2023
Best Road Safety Platform
UK Tech Awards 2022
Best Use of Open Data (beating three government-backed entries)
Transport Research Board
Cited in the 2021 paper on community hazard reporting
Your Junction, The Register
The dataset grows only because 47,800 motorists each year choose to report what they see at a junction — a near miss, a confused give-way, a confusing filter lane. Every confirmed report moves a score, sharpens a warning, and tells the next driver what you already know.
Report a BlackSpot Near YouCross-checked against STATS19 within 48 hours. Your name is never published.