Modelling the heterogeneity in preferences of subway passengers utilizing smart card data from Beijing

There is considerable heterogeneity in the travel behavior of subway passengers. The growing availability of large-scale smart card data offers the valuable opportunity to observe the major share of the passengers over a long period of time to uncover this heterogeneity in travel behavior and devise targeted policies for peak spreading. This study analyzes 7.92 million trip records from Beijing’s subway system—spanning 24 lines and 331 stations—to identify distinct passenger groups based on travel frequency, temporal patterns, and spatial characteristics. Four distinct passenger clusters are identified, and class-specific mode choice models (between subway and bus) are estimated to derive each group’s value of travel time (VOT) and value of travel time reliability (VOR). Results indicate that commuters who take the subway frequently have a higher willingness to pay to improve trip reliability. This indicates that measures like low fare incentives may not be effective in changing the travel behavior of this group, while they are the main target group for alleviating the subway congestion during peak-hours. Therefore, targeted incentives based on the travel characteristics and transport preferences of different passengers could better improve the effectiveness of subway congestion management.