About MobilityMetric

Every signalized intersection in a US city, extracted from OpenStreetMap and cleaned into one record per real junction — 225 cities across all 50 states and DC. This page runs a live 1,000-record sample; the full dataset is maintained continuously.

How the data is built

1. Pull traffic-signal nodes from OpenStreetMap for the area.
2. Cluster and deduplicate overlapping poles into one point per real intersection.
3. Classify each by signal type — vehicular, pedestrian, both, HAWK beacon, or level crossing.
4. Enrich with per-arm approach bearings, speed limits, and street names.
5. Fuse with official city open-data where it exists, and score confidence.
6. Assign a stable, location-based ID that survives data refreshes.

Accuracy vs. published city figures

What each record carries

Stable ID, coordinates, signal type, approach count with per-arm bearings, speed limit (km/h & mph), street names, ADA accessibility flags, OSM-vs-official source & confidence, and last-edited date — exportable as GeoJSON or CSV.

Built for

V2X / C-ITS RSU placement and SPaT/MAP inventories, traffic simulation (SUMO, VISSIM), HD maps, routing, and pedestrian-accessibility audits.