Heat and mobility: Machine learning perspectives on bike-sharing resilience in Shanghai

Global climate change has increased extreme heat events, impacting urban mobility, particularly bike-sharing. This study examines the urban mobility resilience (UMR) of bike-sharing to extreme heat in Shanghai's urban center, analyzing how the built environment and other factors influence UMR through machine learning. The model has an explanatory power of 73.5% and 63.7% on weekdays and weekends. Results indicate that extreme heat has a stronger effect on weekends. Key factors like development intensity, public transportation accessibility, and functional diversity positively affect resilience, with building density increases from 0 to 0.3 promoting UMR by nearly 0.1 (out of 1). Proximity to metro stations (within 1,000-1,500 m) also promotes resilience. The reasonable aggregation of socio-economic factors can also effectively enhance resilience. However, greening and road density sometimes play a negative role. The research provides more references for urban planners and managers to customize strategies to enhance UMR in the context of climate change.