Effects of local, network and systemic dependence on urban development

Identifying the driving factors of urban development and revealing its growth path is an important step toward better understanding the dynamic process of urban evolution. We proposed a theoretical framework of the mechanism of urban development, taking wealth production as an example from the aspects of local, network and systemic dependence to characterize the evolution of cities by temporal persistent trends, network effects and comparative advantages. We first present a detailed description of network dependence and systemic dependence variables to distinguish the differences between hierarchical features of networks based on absolute quantities and the relative performance of individual cities within urban systems. Regression analysis results show that urban economic development relies on its previous performance and presents the characteristic of path dependence. The positive effects of network and systemic variables on urban economic production are significant, especially the closeness centrality and GDP SAMIs ( scale-adjusted metropolitan indicator, deriving from urban expected scaling), emphasizing the indispensable impacts of network effects and systemic relative performance on urban economic development. A machine learning algorithm is applied to simulate urban economic development based on the above factors and achieving great performance, which validate the potential and effectiveness of modeling urban development from local, geographical and systemic aspects. Our findings aid in gaining a better understanding of the implications for urban development and shedding light on a more profound understanding of urban evolution temporally, geographically and systemically.