Toward Identifying the Urban Community Structure from Population Flow and Public Services Distribution

Qinghe Liu,Zhicheng Liu,Yinfei Xu,Weiting Xiong,Junyan Yang,Qiao Wang

DOI: https://doi.org/10.1109/BigData50022.2020.9378225

2020-01-01

Abstract:Identifying the urban community structure is important for promoting smart city. In contrast to community detection in social network, the urban community construction involves both the information of population flow and the distribution of public services. However, few methods take into account the two factors simultaneously in a reasonable manner. This causes several problems such as isolated nodes and geographical fragmentation in urban planning. To address the challenges of concerning the population flow and public services distribution in an unified framework, we formulate the identification problem of urban community structure as graph modularity optimization with constraints on the public services distribution. The combination of hierarchical clustering and node level adjustment strategies is applied, to capture the required information of community structure. In our framework, The attributed graph regards the geographical zones as nodes, the interaction between geographical zones as edges, the distribution of public services as attributes. Moreover, We evaluate our approach by using both the synthetic data set and the real-world data set. Results show the efficacy of the proposed method on identifying the urban community structure.