{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreidq3wgdfz4potg6addzsojac2zue2oyud5utzv4exubisy7o3ridy",
    "uri": "at://did:plc:4tuge3k3comfj4nfvqnwkemn/app.bsky.feed.post/3mluw7xno7sg2"
  },
  "path": "/user/Arjun%20kp/diary/408673",
  "publishedAt": "2026-05-14T18:23:01.000Z",
  "site": "https://www.openstreetmap.org",
  "textContent": "## Contributing commercial vehicle GPS traces from Kerala — a routing approach\n\nI’ve been working on a method to convert commercial vehicle telematics data into useful GPX traces for OSM contribution in Kerala.\n\n### The problem\n\nTelematics data is segment-based — each record has a start coordinate, end coordinate, timestamp and distance, but no continuous GPS track in between. Uploading these directly produces straight lines which aren’t useful for mapping.\n\n### The solution\n\nI set up a local **OSRM** instance using the Kerala road extract from Geofabrik, then route-matched each segment to the actual road network. This produces GPX traces with thousands of road-following points instead of straight lines.\n\n### Results\n\nFrom two months of data covering **Thrissur, Irinjalakuda, Chalakudy, Kodungallur** and surrounding areas:\n\n  * 141 road-snapped segments uploaded as GPX traces\n  * 19 high-priority segments flagged as possible unmapped roads\n  * These will be cross-checked against aerial imagery in JOSM\n\n\n\n### Next steps\n\nReviewing the unmapped road candidates in JOSM against Bing aerial imagery. More vehicle data from the same region will be processed and contributed regularly.\n\nIf anyone in the Kerala OSM community has experience with similar data or wants to collaborate on reviewing unmapped road candidates, feel free to reach out.\n\n* * *\n\n_Tools used: Python, OSRM (self-hosted), gpxpy_",
  "title": "Contributing commercial vehicle GPS traces from Kerala — a routing approach"
}