Map Matching API

The Map Matching API takes raw GPS coordinates and aligns them to the actual road network. Common use cases include fleet tracking and activity replay, cleaning up GPX files from cycling or running apps, verifying that delivery vehicles followed planned routes, and reconstructing historical trips for analytics dashboards.

GPS receivers have a positional error of 5 to 15 metres under open sky, and much more in urban canyons or under tree cover. Raw coordinates often appear off-road, on wrong lanes, or zigzagging across a junction. Map matching applies a Hidden Markov Model to the road network to find the statistically most likely path, producing clean, road-aligned geometries suitable for display and analysis.

The API accepts an ordered array of GPS coordinates as latitude and longitude pairs, optionally with timestamps and bearing values. Timestamps improve match quality on roads with parallel carriageways. The minimum is 2 coordinate pairs. There is no hard upper limit on trace length.

Each matched segment returns a confidence value between 0 and 1. A score above 0.9 indicates a reliable match. Scores below 0.7 typically mean GPS signal was lost, the trace passed through an unmapped area, or the vehicle was off-road. Use low-confidence segments to flag data quality issues or trigger manual review in your pipeline.

Yes. The API returns matched results in milliseconds for typical trace lengths, making it suitable for live vehicle tracking updates. For high-frequency telemetry (more than 1 point per second), batch coordinates into 30 to 60 second windows before sending to keep request volume manageable.

No. Routing calculates a route between two points. Map matching reconstructs the path that was actually taken from recorded GPS data. Use the Directions API when you want to plan a future route. Use the Map Matching API when you want to understand or visualize a route that has already been driven.