Explore city dynamics from the communication network between base stations

Yanghong Zhou, Shijie Zhou, Jiaqing Luo

Abstract: With the continuing urbanisation of the world's population, the risk of economic, health and environmental disaster is also increasing. There is an urgent demand for urban planners and governments to understand how people use the city's infrastructure. Recently, as cell phones become more popular, mobile phone data provides an unlimited source of information about the city. In this paper, we build a social network between different base stations and analyse how the places are connected by using a four-day mobile phone dataset, which consists of 17,598,000 records. Particularly, we construct communication network between different parts of the city and analyse the degree distribution of the network and the homophily of the administrative areas. We then use Louvain method to detect the community structure of the network on weekdays and weekends. Our results show that the in-degree and outdegree distribution of the network exhibit a heavy-tailed form and the assortativity by administrative area is strong. The communities found on weekdays and weekends are both spatially contiguous, which implies that whether at work or not, people tend to communicate within a spatially enclosed community.