Nonparametric Regression Analysis of Cyclist Waiting Times across Three Behavioral Typologies

This paper seeks to predict the average waiting time, defined as the time spent moving at 1 ms−1 or less, of urban bicyclists during rush hours while performing different maneuvers at intersections. Individual predictive models are built for the three cyclist typologies previously identified on a large database of GPS traces recorded in the city of Bologna, Italy. Individual models are built for the three cyclist typologies and bootstrapping has confirmed the validity and robustness of the results. The results allow the integration of waiting times in route choice models for cyclists, thus improving the rational bases by which cyclists makes their decisions. Moreover, the modeling allows transportation engineers to understand how different cyclist typologies perceive different variables that affect their waiting times. Future work should focus on testing the model transferability to other case studies.