A GPS-based bicycle route choice model for San Francisco, California

摘要: AbstractRecognizing the environmental and health benefits of cycling, cities around world are promoting use bicycle for everyday transportation, but with limited information about preferences cyclists effectiveness investments in infrastructure. To better understand decision-making cyclists, we estimated a route choice model GPS data collected from smartphone users San Francisco. Traces were automatically filtered activities mode transfers, matched to network model. Alternatives extracted using repeated shortest path searches which both link attributes generalized cost coefficients randomized. The prior distribution was calibrated only network. A Path Size Multinomial Logit revealed that lanes preferred other facility types, especially by infrequent cyclists. Steep slopes disfavored, women during commutes. Other negative inc...