From Substitution to Complement? Uncovering the Evolving Interplay between Ride-hailing Services and Public Transit

TL;DR

This study reassesses how ride-hailing services (TNCs) interact with public transit (PT) in a mature market by analyzing 16.9 million TNC trips from 96,716 vehicles in Shanghai (Sept 2022) and classifying trips into four categories: first-mile complementary, last-mile complementary, substitutive, and independent. A data-driven framework pairs passenger-provided location labels with PT station access points and uses Amap-derived PT routes to identify substitutions, while CatBoost models with SHAP and PDP uncover nonlinear determinants of the four ratios. The key findings show nearly equal shares for complementary (FCR/LCR) and substitutive (DSR/ASR) trips (about 9%), contrasting with earlier work and suggesting a shift toward mixed modes in saturated markets; nonlinear effects emerge, notably distance to metro stations and bus-stop density, and distinct impacts by metro-station type (single-line vs multi-line hubs). The results yield actionable policy implications, advocating integration and feeder services in underserved areas, enhanced express connections at regional hubs, and a nuanced understanding of how built environment features shape TNC–PT dynamics. The study demonstrates how large-scale, label-based trip data combined with non-linear ML tools can robustly quantify and interpret multi-mode interactions for urban transport planning.

Abstract

The literature on transportation network companies (TNCs), also known as ride-hailing services, has often characterized these service providers as predominantly substitutive to public transit (PT). However, as TNC markets expand and mature, the complementary and substitutive relationships with PT may shift. To explore whether such a transformation is occurring, this study collected travel data from 96,716 ride-hailing vehicles during September 2022 in Shanghai, a city characterized by an increasingly saturated TNC market. An enhanced data-driven framework is proposed to classify TNC-PT relationships into four types: first-mile complementary, last-mile complementary, substitutive, and independent. Our findings indicate comparable ratios of complementary trips (9.22%) and substitutive trips (9.06%), contrasting sharply with the findings of prior studies. Furthermore, to examine the nonlinear impact of various influential factors on these ratios, a machine learning method integrating categorical boosting (CatBoost) and Shapley additive explanations (SHAP) is proposed. The results show significant nonlinear effects in some variables, including the distance to the nearest metro station and the density of bus stops. Moreover, metro hubs and regular single-line stations exhibit distinct effects on first- or last-mile complementary ratios. These ratios' relation to the distance to single-line stations shows an inverted U-shaped pattern, with effects rising sharply within 1.5 km, remaining at the peak between 1.5 and 3 km, and then declining as the distance increases to about 15 km.

Figures (13)

• Figure 1: TNC trip classifications.
• Figure 2: Spatial distribution of ride-hailing trips and public transit stations in Shanghai.
• Figure 3: Framework of recognizing TNC-PT relationships.
• Figure 4: Statistics of complementary and substitutive trips.
• Figure 5: Temporal distribution of complementary and substitutive trips and ratios.