Urban Spatial Interactive Network Construction and Analysis: A Novel Data-Driven Approach

In today's society, the urban spatial interaction network has become an indispensable part of our lives since it can provide the scientific basis for various public services, and improve the quality of life of residents through in-depth analysis of urban spatial interaction relationships. However, the current research on urban spatial interaction networks faces a number of challenges, including the difficulty of obtaining interaction data between cities, the data for constructing spatial interaction networks are relatively single, and the lack of efficient and effective algorithms/models on the spatial interaction networks among cities. To conquer these challenges, this paper utilizes social media check-in data and flight data as data sources to construct a robust urban spatial interaction network, and conducts node importance ranking, structural hole nodes detection, community detection, and city popularity analysis on this network, and applies the research results to urban planning and policy formulation. Among them, we propose an algorithm for complex network node importance evaluation based on vertex entanglement and use the algorithm to rank nodes in terms of importance. In addition, we utilize three algorithms with specific measurement such as modularity and contour coefficient for single-source and integration of social media check-in data with flight data, so as to achieve better community detection. Experiment results demonstrate that the Louvain algorithm is effective in depicting the community structure because it utilizes an improved modularity calculation method and takes advantage of the hierarchical structure of the network.