Differences in metro travel between holidays and nonholidays: spatiotemporal nonlinear interactions with the built environment

Significant increases in metro ridership during holidays place substantial pressure on metro operations and may pose safety risks. Understanding the spatiotemporal characteristics of holiday ridership and its differences from nonholiday periods is, therefore, essential. Using smartcard data from Wuhan, China, this study compares metro travel patterns during International Workers' Day with those on weekdays and weekends across multiple time periods and applies a combined geographically weighted regression and extreme gradient boosting (GWR-XGBoost) model. Results show that average daily ridership during holidays is 1.16 times that of weekends and 1.23 times that of weekdays, accompanied by more dispersed spatial distributions and longer travel times. The influence and relative importance of built environment factors exhibit pronounced temporal and spatial heterogeneity. Unlike non-holiday periods, holiday ridership is concentrated around tourist attractions and closely related to cultural facilities and companies. These findings offer valuable insights for metro operation management and travel safety during holidays.