Speaker
Description
With the enhancement of intercity travel convenience and the effective sharing of urban public resources, commuting patterns have evolved beyond single-city boundaries, increasingly reflecting inter-regional mobility. Compared to intra-city commuting, intercity commuting is characterized by a larger spatiotemporal scope, longer commuting durations, and a greater reliance on transportation infrastructure. Consequently, the choice of intercity commuting is influenced not only by the physical distance between home and workplace and housing costs, but also by factors such as the level of economic development, the availability of public services, and the configuration of intercity transportation networks. Recent research has indicated that the built environment exerts a non-linear and spatially heterogeneous effect on intercity commuting distribution, revealing the complex interactions between these factors (Li and Niu, 2022). However, few studies have managed to address both two characteristics simultaneously.
The study focuses on Nanjing and Ma'anshan as case studies, two cities with a high degree of urban integration in China. The area is divided into 1,756 traffic analysis zones, with data from one month of cellphone signaling analyzed. Preliminary results reveal that, there is an uneven distribution of intercity commuting flows, with north-south commuting significantly surpassing east-west commuting in volume and a pronounced ‘center-to-center’ spatial distribution pattern. To further explore the impact of the built environment on intercity commuting, the study employs an interpretable spatial machine learning framework that combines the Geographically Weighted Random Forest (GWRF) model with SHapley Additive exPlanations (SHAP) analysis, uncovering the complex spatial factors influencing intercity commuting patterns.
The GWRF model effectively combines the spatial analysis capabilities of Geographically Weighted Regression (GWR) model with the predictive power of Random Forest (RF) model to better reveal the potential spatial heterogeneity and nonlinear effects of the variables, while SHAP enhances interpretability(Wu, Zhang and Xiang, 2024). Preliminary results show that employment density and public transportation coverage are the primary factors influencing the attractiveness of intercity commuting (with importance values of 16.34% and 14.47%, respectively). The workplaces of intercity commuters typically exhibit spatial agglomeration characteristics, with commuting attractiveness increasing sharply when employment density exceeds 4,500 jobs/km², demonstrating the nonlinear positive effect of employment density on commuting. Furthermore, since intercity commuters typically face longer travel distance, their demand for convenient transit transfers is higher (Bouzouina et al., 2021). However, the study also finds that the attractiveness of commuting plateaus when transit route density exceeds 106 routes/km², indicating a saturation effect of transit coverage and accessibility on intercity commuting. From the commuting generation perspective, land prices and educational resources emerge are the two critical influencing factors (with importance values of 19.56% and 13.62%, respectively). This suggests that when housing prices in large cities exceed the affordability threshold for intercity commuters, they tend to choose residence in neighboring cities, aligning with findings from previous studies (Chen, Ge and Pan, 2021). The strong positive relationship between educational resources and intercity commuting is attributed to the abundance of educational resources in Nanjing and proximity-based enrollment policies that attract residents from nearby cities.
Based on the above findings, the study proposes differentiated optimization strategies to enhance urban morphology and transportation layout, thereby improving intercity commuting efficiency. The results provide a theoretical foundation for optimizing urban design and transportation networks, as well as practical guidance for urban planning and transportation policy, particularly in the context of promoting the integration of urban agglomerations and metropolitan areas, offering significant practical value.
References
[1]Zhipeng, L ,Xinyi, N. (2022) ‘Exploring Spatial Nonstationarity in Determinants of Intercity Commuting Flows: A Case Study of Suzhou–Shanghai, China’, ISPRS International Journal of Geo-Information, 11(6), pp. 335-335.
[2]Dongyu, W ,Yingheng, Z ,Qiaojun, X. (2024) ‘Geographically weighted random forests for macro-level crash frequency prediction’, Accident Analysis and Prevention, 194, pp. 107370-107370.
[3]Louafi, B ,Ioannis, B ,Patrick, B , et al. (2021) ‘Renters vs owners: The impact of accessibility on residential location choice. Evidence from Lyon urban area, France (1999–2013)’, Transport Policy, 109, pp. 72-84.
[4]Tao, C ,Yanbo, G ,Haixiao, P . (2021) ‘Car ownership and commuting mode of the “original” residents in a high-density city center’, Journal of Transport and Land Use, 14(1), pp. 105-124.
| Keywords | Intercity Commuting; Built Environment; GWRF model; Nonlinear Effects; Spatial Heterogeneity |
|---|---|
| Best Congress Paper Award | No |
