A global framework to estimate urban spatial cycling patterns based on crowdsourced data

A team of researchers from the University of Helsinki and Aalto University has published a novel framework that leverages AI and crowdsourced data to create a global picture of urban cycling. The study, 'A global framework to estimate urban spatial cycling patterns based on crowdsourced data,' addresses a critical data gap in sustainable mobility planning by repurposing the openly accessible Strava Global Heatmap. Traditionally, cycling research has been hampered by inconsistent, inaccessible, or non-existent official count data, making large-scale comparative analysis difficult. This new method provides a standardized, low-effort alternative for cities worldwide.

The core technical innovation involves refining the raw Strava heatmap—which shows relative cycling intensity—by weighting it with local population density and, more effectively, with counts of Points of Interest (POIs) like shops and services. When validated against physical cycle counter data from 29 global cities, the POI-weighted model showed strong correlations, particularly in European and North American East Coast cities, with most achieving p > 0.8. The framework's accuracy improves in cities with a higher existing cycling modal share. This tool enables urban planners and researchers to generate categorical volume estimates and conduct consistent, large-scale analyses, directly supporting data-driven decisions for bike lane investments and network planning where traditional data is lacking.