Comparative Analysis of Usage Patterns and Underlying Determinants for Ride-hailing and Traditional Taxi Services: A Chicago Case Study

Abstract

As app-based ride-hailing (or e-hailing) services have achieved great success worldwide, the traditional taxi industry faces an unprecedented crisis. To gain a deeper understanding of these two distinct modes that both offer door-to-door mobility services, in this study, we use Chicago as a case to comparatively analyze the travel patterns of the app-based ride-hailing and traditional taxi services and understand how socio-demographic and urban land use attributes affect the usage of two services in temporal and spatial dimensions. We employ the local regression models named geographically and temporally weighted regression (GTWR) models that can capture the spatiotemporal nonstationarity between variables as the primary analysis tool. Four GTWR models that predict the average hourly taxi or ride-hailing trip volume on weekdays or weekends for each community area are carefully constructed with R^2 all above 0.98. The nonstationarity test shows all variables exhibit extra local variations, indicating the necessity of using the GTWR models to explore the attributes’ spatiotemporal impacts. Spatial error models (SEM) and ordinary least square (OLS) models, global regression models, are also built as comparisons and benchmark references. The differences between global and local regression models are illustrated through clustering analysis. To compare and contrast the usage patterns of ride-hailing and taxi services, we analyze the temporal-spatially varying coefficients of GTWR models. Interpretations are provided for our interesting findings, including the strong trip generation power of the population in the South Side area, the attitude of white people towards taxi services is positive but ambiguous towards ride-hailing services, young people's opposing behaviors towards taxi services on weekdays and weekends, and the opposite relationship between transit and taxi or ride-hailing in certain areas. Based on these key findings, we offer some planning and operational suggestions, such as developing mobile apps, implementing price-related strategies, strengthening partnerships with local businesses, building multi-modal transportation, clarifying customer segmentation, providing auxiliary services, etc.