Understanding the determinants of spatial-temporal mobility patterns based on multi-source heterogeneous data

With the advance of intelligent transportation systems (ITSs) and data acquisition system (DAS), it is possible to explore the determinants of urban spatial-temporal mobility patterns using multi-source heterogeneous data. This study aims to use the points-of-interests (POIs) data, house-price data and floating car data to identify the factors influencing urban mobility. Within a scale of 0.5 km grid, trip production and attraction were stratified according to the traveling intensity, and the critical information related to economy, intermodal connection, land use and time were also obtained through the multisource data. The experiment results from an ordinal logistic regression (OLR) analysis show that average house price has a dominating and positive effect on the traveling intensity for both trip production and attraction, followed by land use factors including finance, healthcare, residence, living service, hotel, cultural and educational service, government, corporation, catering, as well as leisure and sports service. However, the effect of scenic spots is found significant only on trip attraction. In addition, shopping is found to insignificantly affect the traveling intensity for both trip production and attraction. Unexpectedly, time factors also have diverse impacts. These findings are expected to help better understand the relationship between urban mobility and built environment factors, offering useful insights on urban planning.