Exploring Urban Spatial-temporal Patterns via Large-scale Vehicle Travel Data: The Role of Geographical Attributes and Traveler Characteristics

Studying the spatial-temporal patterns of vehicle travel helps to get more insights about the urban transportation, and contributes to multiple downstream management or planning tasks in urban system. When existing researches usually work on of taxi records or highway data which lead to bias observations, we conduct a comprehensive exploration of vehicular spatial-temporal patterns using large-scale checkpoints data of all kinds of vehicles in urban road networks. Specifically, we extract different patterns with OD-tensor decomposition and discover their geographical localized modes. We investigate the influences of the urban geographical attributes on these OD travel patterns, pointing out the most important factors. We further cluster the individual vehicular travelers into different types of groups, and show their distribution differences under different spatial-temporal patterns. The obtained results can provide decisions support for transportation infrastructure construction and urban functions planning.