Factors Affecting Urban Sustainable Development by Metro Using the GCN–MGWR model

This study investigates the relationship between urban subway passenger flow and land use intensity, proposing an innovative hybrid model that combines graph convolutional networks and multiscale geographically weighted regression (GMGWR). This model addresses the limitations of traditional methods in handling nonlinearity and spatial heterogeneity. Using metro data from Chengdu, Sichuan, China, this study analyzes the effects of various land use types on metro passenger flow during different time periods, revealing the spatial and temporal dynamics of land use on the urban rail transit system. The results indicate that land use characteristics are key determinants of urban rail transit passenger flow and that the effects of land use intensity on metro passenger flow exhibit dynamic characteristics that change with time and space. The innovation of this study lies in integrating machine learning and spatial econometrics methods. The proposed GMGWR model provides a more accurate representation of the complex nonlinear relationship between land use and metro passenger flow, offering urban transportation planners valuable strategies to enhance public transportation systems. By strategically planning land use around metro stations and promoting transit-oriented development policies, it is possible to create livable, pedestrian-friendly communities that foster green, sustainable urban growth.