Identifying Individual Activity Patterns from Mobile Phone Tracking Data

Human mobility exploration through data mining gains many benefits from massive digital data sources. Geolocation data of mobile phones involves users' spatio-temporal geographic information, but it does not include explicit labels of activities. This paper investigates individual activity patterns based on one-month mobile phone tracking data in the Paris region of France. The semi-definitive activity labels, including the primary anchor places and secondary activity places, are first extracted. The criteria of the cumulative presence duration and visiting frequency in activity locations over the study period are used to identify these places. To find individual activity patterns, characteristics such as activity frequency and activity duration related to the activity places are then investigated. Individual neighbor activity space is also measured around identified home and work places. Based on the individual activity features, we statistically analyze a set of activity patterns for all samples in our case study.