EconPapers    
Economics at your fingertips  
 

Investigating relationships between ridesourcing and public transit using big data analysis and nonlinear machine learning: A case study of Shanghai, China

Xinghua Liu, Qian Ye, Ye Li, Kaidi Yang and Xuan Shao

Transportation Research Part A: Policy and Practice, 2025, vol. 192, issue C

Abstract: Ridesourcing has transformed the landscape of passenger transportation systems in many cities worldwide, but whether it competes with or complements public transport (PT) is still debated, and the literature is limited. Therefore, this study aims to address this knowledge gap by measuring the relationships between the two systems and examining their determinants using a multisource big data analysis and nonlinear machine learning approach, with Shanghai, China, as the study case. First, we used the observed ridesourcing data in Shanghai to compute the fastest PT alternative for each ridesourcing trip based on the Amap open platform and subsequently compared the travel patterns (i.e., distance, duration, and generalized cost) of the two systems. Second, we propose a technical framework that considers the spatiotemporal availability and generalized cost acceptability of PT services, as well as the inclusivity of ridesourcing services, to accurately classify and identify the relationship between ridesourcing and PT systems. Finally, we explored the importance of four types of determinants, namely, ridesourcing characteristics, PT service, built environment, and weather, and their nonlinear effects on different relationships based on extreme gradient boosting and Shapley additive explanations. Our results show that the fastest PT alternative involves an average travel distance, generalized travel time, and generalized cost that are 1.16, 2.13, and 1.15 times greater, respectively, than those of ridesourcing. Competitive trips account for 36% of urban areas but only 16% in the suburbs. Furthermore, more than 70% and 10% of the ridesourcing trips in suburban areas are used to complement and integrate PT, respectively. The nonlinear machine learning framework identified the top three determinants of integration as travel cost, distance to the CBD, and travel time. Notably, determinants such as the distance to the CBD and temperature have nonlinear effects on these relationships. These findings offer valuable insights for designing multimodal transportation options that integrate the benefits of ridesourcing and PT.

Keywords: Shared mobility; Ridesourcing; Public transit; Machine learning; Big data; Shanghai (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0965856424003872
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:transa:v:192:y:2025:i:c:s0965856424003872

Ordering information: This journal article can be ordered from
http://www.elsevier.com/wps/find/supportfaq.cws_home/regional
https://shop.elsevie ... _01_ooc_1&version=01

DOI: 10.1016/j.tra.2024.104339

Access Statistics for this article

Transportation Research Part A: Policy and Practice is currently edited by John (J.M.) Rose

More articles in Transportation Research Part A: Policy and Practice from Elsevier
Bibliographic data for series maintained by Catherine Liu ().

 
Page updated 2025-03-19
Handle: RePEc:eee:transa:v:192:y:2025:i:c:s0965856424003872