The uncertain geographic context problem in identifying activity centers using mobile phone positioning data and point of interest data

Abstract

People aggregate at different areas in different times of the day, thus forming different activity centers. The identification of activity centers faces the uncertain geographic context problem (UGCoP) because people go to different places to conduct different activities, and also go to the same place for carrying out different activities in different times of the day. In this paper, we employ two kinds of novel dynamic data, namely mobile phone positioning data and Point of Interest (POI) data to identify the activity centers in a city in China. Then mobile phone positioning data is utilized to identify the activity centers in different times of a working day, and POI data are used to show the activity density variations at these activity centers to explain the temporal dynamics of geographic context. We find that mobile phone positioning data and POI data as two kinds of spatial-temporal data demonstrate people’s activity patterns from different perspectives. Mobile phone positioning data provide a proxy to represent the activity density variations. POI data can be used to identify activity centers of different categories. These two kinds of data can be integrated to identify the activity centers and clarify the UGCoP.