Understanding taxi ridership with spatial spillover effects and temporal dynamics. (June 2022)

• Record Type: Journal Article
• Authors: Zhu, Pengyu; Huang, Jie; Wang, Jiaoe; Liu, Yu; Li, Jiarong; Wang, Mingshu; Qiang, Wei
• Is Part Of: Cities. Volume 125(2022)
• Publication Date: 2022-06
• Subjects: Taxi demand -- Spatial econometric model -- Spatial autocorrelation -- Big data -- Built environment; City planning -- Periodicals; Urban policy -- Periodicals; 711.4
• DOI: 10.1016/j.cities.2022.103637
• Languages: English
• ISSNs: 0264-2751

Abstract

Abstract: In urban transportation systems, taxis are regarded as flexible, convenient, and time-saving. Taxi demand is affected by various built-environment factors and by the time of the day. Although many studies have investigated correlations between taxi demand and the built environment, the direct and spillover effects of built environment factors on taxi demand have not been examined at a fine spatial scale. To address this gap in the literature, this paper employs spatial econometric models using GPS-tracked taxi trips, mobile signaling data, and points of interest (POIs) to study taxi demand in Beijing at a 1-kilometer square grid resolution. The results show that, in the morning and evening peak hours, road network density has the strongest (positive) direct and indirect impact on taxi ridership. A relationship is also found between public transportation and taxi ridership: bus coverage has positive direct effects and insignificant indirect effects on taxi pick-ups and drop-offs, while subway coverage has negative indirect effects, suggesting that it may absorb taxi demand from surrounding grids. Results also indicate that various built-environment factors affect taxi demand differently at morning and evening peak times. This study reveals the complex nature of taxi ridership and has important implications for policymakers, transport planners, and other stakeholders in megacities around the world. Highlights: Use spatial autoregressive models to investigate the relation between built environment, socioeconomics and taxi ridership. Conduct the analysis for morning and evening peak hours. Built environment characteristics and socioeconomics influence taxi ridership via both local and spillover effects. Offer a framework to incorporate various types of big data, including taxi data, mobile signaling data, points of interest.