Analyzing the spatial-temporal influence of built environment on taxi ridership

Abstract

Taxi is an important component in urban transportation system, which covers wide area and maintains 24h available. Exploring the relationship of taxi and built environment is very important to manage taxi service and improve transportation system. The issue is addressed by capturing the influence of built environment on taxi ridership considering spatial and temporal non-stationarities. The grid cells are developed as analysis units and the global regression model is adopted for preliminary exploring. Then the GWR are implemented considering spatial heterogeneity and the GTWR is used to analyze the spatiotemporal influence of built environment on taxi ridership. An empirical study conducted in Hong Kong Island using one-week taxi’s GPS data demonstrates the effectiveness of the regression models. It’s verified that GWR performs better than OLS, and GTWR outperforms the rest two regression models, indicating both time and space are critical dimensions.