Modeling cyclists' facility choice and its application in bike lane usage forecasting

In some circumstances on streets equipped with new bike facilities, cyclists are not interested in using them. Instead, they continue to use shared spaces with pedestrians or motor vehicles. Thus, simply adding a bike facility does not guarantee that cyclists will switch to using it. Owing to the considerable development of bike facilities, the investigation of facility preference, particularly focusing on facility choice forecast, has become increasingly important. This study developed a model for predicting the facility choice of cyclists between on-street facilities (curb, traffic lane, and bike lane (BL)) and off-street facilities (sidewalks). Initially, the optimal model was selected using Bayesian Model Averaging method. Then, it was validated by both internal and external validations. Apart from the aforementioned factors, several other exogenous variables were also found to be significant predictors of bike facility choice, including the width of traffic lanes, existence of real-time stopping vehicle, type of bike, bus stop existence, and in-group cycling. Analysis of the relative importance of predictors indicated that bus stop existence, effective sidewalk width, and type of bike were the potential predictors. A framework for predicting BL usage, if it is present, was also developed. A test for the predictive performance of the application at a real site was carried out. By comparing predicted and actual BL usage figures, the analysis showed good predictive performance. The results of this study can help developers, planners, and designers to adopt reasonable investment decisions as well as better designs in developing new bike facilities.