Shared e-bikes demand in urban mobility: Temporal heterogeneity, driving factors, and strategic implications

Electric bicycles (e-bikes) are recognised as a sustainable solution for urban travel, addressing increasing travel demand while reducing emissions. Shared e-bikes, with increasingly developed infrastructure and integration with existing urban transportation systems, offer an efficient alternative, especially in short-distance travel. Understanding the variation in shared e-bike usage and its interaction with external factors is important for optimising system performance. Using shared e-bike transaction data from Nanning, this study employs traditional statistical regression (Spatial Lag Model, SLM) and machine learning (Extreme Gradient Boosting, XGBoost) to reveal demand variations across seven distinct time segments. SLMs identify the significance and coefficient variations of explanatory variables, while XGBoost reveals shifts in feature rankings and influence thresholds. Findings from both models highlight the significant influence of public transit and certain facilities on shared e-bike demand, with notable temporal patterns. Results indicate shared e-bikes’ role in facilitating users’ weekday routines, particularly commuting, and supporting leisure activities around areas populated with restaurants and universities during after-work hours. Shared e-bikes also show a positive trend of integration with current transit systems. These findings suggest a deployment focus that varies by time: balancing shared e-bikes around transit stations, residential areas, and office buildings during weekday peak hours, and around restaurant-populated areas and transit stations as the day unfolds; during weekends, prioritising transit stations, universities, densely populated areas, and restaurants. With guided attention to different time slots, the findings will help operators optimise resources and enhance service outcomes.