Transformation of Relationship between Shared-bike and Public Transit in Seoul, Korea: Application of the SHAP Model

Abstract

Public bicycles can be effective alternatives to the massive public transit during external shocks like strikesor epidemics. In addition, compared to other developed countries, the bicycle share of the whole transitsystem is insufficient in Korea. Therefore, this research aimed to offer targeted advice to improve public bikeservices in terms of their interaction with the mature massive public transit system.Firstly, this research used real operating data of public transits and bike-sharing trip data to investigate thespatial distribution of the relationship between bike-sharing usage and public transit ridership, includingsubstitute, integration, and complement. Secondly, the method was used to analyze the difference between2018 and 2019. Thirdly, environmental factors contributing to the heterogeneity of the modal relationshipbetween bike sharing and public transit and the cold spots of bike-sharing usage were detected by themachine learning regression model. It is shown that the integration is spatial heterogeneous according to thedistance, location, and accessibility to the public facilities. Furthermore, the results detected the insufficientarea for the bike-sharing supplement in Seoul in terms of each type of function.These results will stimulate the usage of low-carbon transit modes rather than indulge in energy-consumingautomobiles during the pandemic when passengers are crowd-averse during the pandemic.