Characterizing and Comparing User Location Preference in an Urban Mobile Network

With fast increasing numbers of 3G mobile subscribers and accompanying growth in traffic, a deep understanding of user mobility and traffic usage patterns is vital for network resource allocation and optimization in order to maintain or even improve user experience. In this paper, we report results from a one-week network trace of an urban mobile network with over 200,000 users. We first propose several normalized metrics such as the ratio of active users (RAU) and its z-score (ZLP) either by time or by user group, which allow comparison of location preferences across different user groups. Next, we apply these metrics on an area of 50 km(2), composed of 358 lattices. The results show that the top ranked favorite lattices are located in the west and middle parts of this given area. The k-means clustering of lattices reveals their time sequence feature patterns. Furthermore, we specifically analyze users having different ARPU (Average Revenue Per User) levels. We quantify and compare the degrees of location preferences, and present them numerically and visually. These statistics would be utilized in various practical scenarios for enhancing user experience.