Travel Time Prediction from Sparse Open Data
A new random forest model offers a free, accurate alternative to costly APIs like Google Maps.
Deep Dive
Researchers Geoff Boeing and Yuquan Zhou developed a free, open-source travel time prediction model. It uses a random forest model trained on sparse, publicly available data like naive travel times, turns, and traffic controls. Validated in Los Angeles, it provides a 'middle-ground' technique that balances reasonable accuracy with minimal data, computational, and cost requirements, making metropolitan-scale trip analysis accessible to less-resourced professionals without expensive APIs.
Why It Matters
Enables affordable, large-scale urban planning and accessibility analysis, democratizing tools previously locked behind costly commercial APIs.