Visualizing population mobility from spatiotemporally aggregated mobile phone data via a 3D gradient approach

Population mobility patterns are important for understanding a city's rhythms. With the widespread use of mobile phones, population-based trajectories can be utilized to explore such mobility patterns. However, to protect personal privacy, mobile phone data must be de-identified by data aggregation within each spatiotemporal unit. In data acquired from mobile phones, population mobility features are still implicit in the spatiotemporally aggregated grid data. In this study, based on image-processing techniques, a two-step 3D gradient method is adopted to extract the movement features. The first step is to estimate the initial movement pattern in each spatiotemporal grid, and then to estimate the accumulated movement pattern within a time period around a geographical grid. This method can be applied adaptively to multi-scale spatiotemporal grid data. Using geospatial visualization methods, estimated motion characteristics such as velocity and flow direction can be made intuitive and integrated with other multiscale geospatial data. Furthermore, the correlation between the population mobility pattern and demographic characteristics, such as gender and age groups, can be analyzed with intuitive visualization. The implication of the visualization results can be used for understanding the human dynamics in a city, which can be beneficial for urban planning, transportation management, and socioeconomic development.