Search for a command to run...
Urban areas face multiple challenges stemming from population growth and the limited capacity of transportation systems. Sustainable transport modes, such as cycling, are crucial for addressing congestion, pollution, and accessibility. Bike infrastructure improves overall accessibility but accessibility also depends on actual and perceived safety and comfort. To measure these nuanced aspects, a broader indicator is needed. Bikeability aims to bridge this gap. It can be approached in two different but complementary ways. Either quantitatively through objective measurements or simulations (e.g., type of infrastructure, slope, speed and traffic) or qualitatively. This paper proposes a hybrid, reproducible framework that combines quantitative and qualitative aspects. Using five indicators based on open data (type of infrastructure, slope, speed of motorized traffic, number of lanes and environment), a user survey and machine learning, we propose a bikeability assessment model. Results from our decision tree model show that physical separation is the primary factor defining bikeability, splitting the remainder of the model into a safety-focused branch and a comfort-focused branch. Cyclists prioritize aesthetics and comfort only after basic infrastructure separation guarantees safety. Validated through a case study of Lyon, France, the results show a high correlation between top-rated segments and the most frequented biking spots in the city. This framework provides urban planners with a scalable, low-cost diagnostic tool to prioritize infrastructure investments without relying on expensive sensing campaigns.