Fine-Grained Dynamic Population Mapping Method Based on Large-Scale Sparse Mobile Phone Data

Abstract

The dynamic nature of urban population distribution plays a key role in urban planning, emergency management and public travel information services. Currently, the widespread use of mobile phone data provides the opportunity to support fine-scale population studies. However, the data sparsity problem of mobile phone data has been a huge handicap. To overcome this, we proposed a comprehensive approach to achieve fine-grained dynamic population distribution and high-resolution population map based on large-scale sparse mobile phone data. First, we developed an anchor-point-based trajectory reconstruction method to improve the spatiotemporal granularity of mobile phone trajectories. Then, a rapid and efficient automation population mapping method was proposed with the support of reconstructed high spatiotemporal resolution of human movements. Finally, we analyze spatiotemporal characteristics of population distribution and spatial-temporal interaction of human movement. Using a real mobile phone dataset in the city of Shanghai as a case study, we evaluated the performance of our method. Results indicated that our method improved the precision and reliability of population distribution estimation and could be utilized for quantitatively analyzing the spatiotemporal characteristics of population distribution and migration. We argue that this study is useful for understanding the highly dynamic human movement states and supporting advanced urban applications.