Study on Interactive Network among Chinese Cities Based on the Check-in Dataset

刘瑜,隋正伟,邬伦
DOI: https://doi.org/10.7702/dlydlxxkx20130601
2013-01-01
Abstract:In the era of big data,large volumes of geo-tagged check-in records,collected from on-line location based services such as Sina Micro-blog,Jiepang,and Renren Places,covering millions of user,provide a promising approach for investigating large-scale individuals′trajectory patterns as well as spatial interaction characteristics among cities.In this paper,applying complex network theory and methods,the statistics and community structure of an undirected weighted network are revealed,which includes 370cities in China.The network is constructed according to large amounts of users′trajectories between cities.Each edge in the network represents the interaction strength between two cities.Using complex network analyzing features,such as degrees,diameter,clustering coefficient,average path length,the result shows that the Chinese cities′interactive network exhibits a small-world feature.Based on the community detection method——multi-level method,the modularity is used to compare the quality in the process of hierarchical community structure detection.The results show that the regional structure characteristics of the interactive network among Chinese cities and the inter-urban interaction strengths show a power-law distribution.The Chinese city system can be divided into 8national level areas and 21provincial level areas and the divided area boundary coincide with the administrative boundary very well.
What problem does this paper attempt to address?