Investigating user preferences for dockless bike- and electric bike-sharing through tracking usage patterns

The integration of electric bikes has significantly boosted the popularity of shared micro-mobility. To promote the coordinated development of dockless bike-sharing (DBS) and electric bike-sharing (EBS), it is crucial to analyze the mechanisms influencing user preferences. However, capturing accurate usage patterns of users remains a challenge, hindering the optimization of shared micro-mobility services. Using one month of shared cycling order data from Kunming in 2022, this study tracks user travel patterns and categorizes them into three types: DBS-dominant, balanced, and EBS-dominant. To investigate the underlying mechanisms influencing these preferences, the study initially applies HDBSCAN clustering to identify users' frequent travel locations. A weighted Gradient Boosting Decision Trees (GBDT) model is employed to reveal the nonlinear relationship between explanatory variables and user preference types. The model considers factors from the perspectives of travel characteristics, built environment, and shared infrastructure systems. Results indicate that travel characteristics and the built environment significantly affect users' travel preferences. DBS-dominant users prefer short-distance, high-frequency trips, particularly in the Central Business District (CBD) and areas with complex road conditions. In contrast, EBS-dominant users favor long-distance travel and prolonged use, particularly in areas farther from the CBD. Balanced users exhibit flexibility, switching between DBS and EBS based on specific needs and conditions to maximize convenience. Targeted policy measures are proposed for various user groups to improve travel services and support the integrated development of the DBS and EBS systems. This study not only provides scientific decision-making support for shared mobility services but also assists market operators in refining their offerings.