Xu, Chun2022-09-082022-09-082022-08-15https://hdl.handle.net/1813/111573Inclusivity of walking and biking is crucial for cities to adapt to climate change and towards a sustainable future. The bike-friendliness of neighborhoods is determined by attributes of the built environment. Using thousands of sharing-bike trajectories in Ithaca, this study created a bike-friendly index which revealed how cyclists’ preferences for routes deviate from the shortest path by calculation. To explain the deviation, 18 environmental attributes across six dimensions (namely topography, bicycle infrastructure, traffic, comfort, attractive, and land use) are measured using geospatial data and computer vision techniques. Generally, gentle slopes, wider roads, attractive and comfortable urban design lead to a bike-friendly street. The existence of bikeways and bus stops, as well as low traffic volume, do not contribute to the street’s bike-friendliness in Ithaca. The results furthered our understanding of the role of the built environment in affecting biking route choices with mobility big data and artificial intelligence techniques.enAttribution-NonCommercial-NoDerivatives 4.0 Internationalbike-friendlysustainable transportationbuilt environmentgeospatial analysisartificial intelligencecase study