Statistical analysis and prediction of regional bus passenger flows

Abstract

Urban bus transit is a major mode of transportation in modern cities and plays an important role in mitigating the traffic pressure in urban road networks. We used smartcard data collected by three-million bus passengers in Shenzhen, a major southern city of China, to study the statistical properties and dynamics of bus passenger flows. In this study, the recorded passenger flows were cross-grained into each 1 km × 1 km square grids to avoid large flow variations at a single bus station. The temporal and spatial patterns of passenger flows were analyzed and a machine learning-based model for passenger flow prediction was generated.