Abstract
Understanding human mobility is crucial for every aspect of daily life and the functioning of cities. Advanced by sensor technology and the big data economy, a highly influential body of research and applications on human mobility is driven by analyses of massive human location datasets, such as social media data and spending data. New data is emerging as rapidly as evolutionary technologies. Mobile app data is relatively new and has become available only in the recent decade. The derived data products are similar to those mainstreaming existing ones, mainly in trip-activity chains, counts, flow matrices, and derived indicators. However, the data bias varies across areas, periods and policy restrictions, requiring tailored data processing and validation solutions, which are not fully transparently discussed. This study contributes as a handbook for processing similar types of location points data, detailing engineering workflow and multi-stage validation techniques. Second, we present insights into the limitations and potential of data applications that tolerate the inevitable data bias. Finally, open trajectory and matrix data are shared for research purposes. The team will keep updating the methodology and results with the latest developments on GitHub.
Authors: Chen Zhong, Nilufer Sari Aslam, Yikang Wang, Zhengzi Zhou and Adham Enaya
Download CASA Working Paper 240 (file size 3 MB, file format pdf)