Evaluating spatial coupling between employment, population density, and metro stations in Shenzhen: A network accessibility perspective

Transit-Oriented Development (TOD) is increasingly recognized as an effective urban planning strategy, aiming for the harmonious development of urban spatial structures and rail transit systems, thus fostering sustainability. This study addresses the limitations of existing research in quantifying the relationship between rail transit and the spatial distribution of residential and employment populations. To bridge these gaps, this study develops an Accessibility of Metro Stations (AMS) index to examine how metro accessibility influences employment and population densities, offering recommendations for optimizing TOD strategies. Methodologically, the study employs social network analysis to develop the AMS index, measuring the connectivity of metro stations to surrounding areas. Cellular signaling data from Shenzhen provides spatial accuracy, supporting a spatial autocorrelation analysis of the coupling between AMS and both employment and population densities. Findings reveal a moderate correlation and a significant positive spatial correlation between AMS and these densities, with a more robust association observed in the resident population. Intriguingly, only 41% of metro stations exhibit a coupling relationship with the resident population, and 33% with the employed population. Based on these insights, a typology of coupling relationships is proposed, linking metro station accessibility to population and employment densities. This typology is instrumental in informing targeted urban management policies and guiding practical TOD implementations, fostering sustainable urban development and promoting long-term environmental, economic, and social resilience. • Metro stations in the central area attract more employed population. • There is a moderate correlation between metro stations and urban development. • Metro has a more substantial role in guiding the resident population. • There are four categories of agglomerations in the spatial coupling.