Identifying London’s near-miss clusters to predict crashes (Phase 1)

Transport for London (TfL) wanted to understand whether near-miss events could act as early indicators of crashes, intervening in high-risk areas before fatalities occurred.

This approach is informed by Kent County Council’s Crash Remedial Programme (CRM) and near-miss analysis.

Specifically, TfL wanted to:

• Identify clusters where near-misses coincide with crash occurrences.
• Evaluate the extent to which Compass near-miss data overlaps with TfL crash records.
• Test whether historical near-miss patterns can be predictive of future crash hotspots.
• Create a data-driven dashboard to support proactive safety interventions.

Compass' Data Science team developed a custom dashboard to pinpoint clusters across the entire network (macro-analysis). Then, Compass' Road Intelligence platform was used to analyse risky road segments (micro-analysis).

Watch the video to learn more: