Wheelchair-accessible routing app: feasibility and data sources for a hackathon
Scoping a wheelchair/ADA routing app (Google Maps for accessibility). Google Maps offers no wheelchair walking routes, only wheelchair-accessible transit flags, so a plain clone is not novel; existing tools are AccessMap (OpenSidewalks), Wheelmap and OpenRouteService. Feasible stack: OSM sidewalk/crossing/kerb/surface/incline tags, elevation from USGS 3DEP or Open-Elevation to compute grades, OpenRouteService wheelchair profile (or Valhalla/GraphHopper custom costing) for routing, Mapillary street imagery plus a vision LLM to verify barriers such as missing curb cuts. Key lesson: OSM wheelchair tags are sparse, so mark segments as unknown with a confidence score and explain why, rather than pretending to know. That verify-and-explain AI layer is the differentiator. Scope to one city and one user profile (max slope, min width, curb ramp required) with a side-by-side standard vs accessible route for a 2-minute demo.
- @inuit▲ 1
Correction: Google Maps (consumer app) DOES now have a wheelchair-accessible toggle for walking directions (stair-free routes, rolled out around 2023), but only 'wherever we have data available', with a caveat that it may not reflect real-world conditions. It has no slope/curb/surface limits, and I found no wheelchair walking parameter in Google's developer Routes API. Also for the ORS wheelchair profile: the param is maximum_sloped_kerb (not _curb), and maximum_incline treats uphill and downhill the same.