Inferring Social Influence in Transport Mode Choice Using Mobile Phone CDR Data

Abstract

Previous chapters have shown how mobile network data can be used to infer travel generation and trip distribution. We continue to explore further in this chapter into how to utilize the mobile network data in making inferences about transport mode choice and its social influence. This chapter focuses on social influence in terms of ego-network effect in commuting mode choice, for which a longitudinal mobile phone data that includes both location and communication records is investigated. Methods for inferring social tie strength and transport mode as well as a framework for analyzing social influence in transport mode choice are discussed. The findings reveal that a person’s strong relationships are more essential in determining whether or not driving is the person’s preferred mode of transportation, whereas weak ties are more relevant in determining whether or not public transportation is the person’s preferred mode of transportation. The data also shows that social ties that are geographically nearby have a greater influence on commuting mode choice than those that are farther away. In the case of public transportation, accessibility distance is also a deciding factor. As the access distance increases, the percentage of people who use public transportation decreases. Furthermore, the social network has been found to influence commute mode choice, with the likelihood of choosing a given mode increasing as the percentage of social ties who choose that mode increases. The content discussed in this chapter reflects the idea, motivation, and thinking process in our original work done by Phithakkitnukoon et al. (EPJ Data Sci. 2017;6(11); Soc Netw Anal Min. 2016;6(1)).