On human mobility predictability via WLAN logs

Abstract

In this research, we conduct a comprehensive measurement study on the predictability of human mobility with respect to demographic differences. We leverage an extensive WLAN dataset collected on a large university campus. Specifically, our dataset includes over 41 million WLAN entries gathered from over 5,000 students (with demographic information) during a four-month period in 2015. We observed surprising patterns on large increases of long-term mobility entropy by age, and the impact of academic majors on students long-term mobility entropy. The distribution of long-term entropy follows a bimodal distribution, which is different from previous studies. We also find that the predictability of students' short-term (daily or weekly) mobility varies on different days of the week and with student gender. Because of the large campus size, our results can mimic people's mobility patterns in metropolitan areas. We also anticipate that our results will provide insight that guides academic administrators' decisions regarding facilities planning, emergency management, etc. on campus.