Investigating Commuters’ Leisure Activity Choice Behavior Based on Metro Smart Card Data

Smart card data collected by automated fare collection systems record massive amounts of travel information. Processing and analyzing these data open new opportunities in travel behavior research. This paper aims to study leisure activities of commuters in urban areas based on smart card data from Nanjing’s metro system. Four relevant indicators related to leisure activity choice behavior are adopted to classify commuters, i.e., activity duration, weekly frequency, time preference, and stability of activity location. Commuters are divided into three categories through cluster analysis; 64% of commuters are found to rarely conduct leisure activity, a preferred activity time and no stable activity location. A Markov model is introduced to predict leisure activity choice behavior for each type of commuter. The results show that leisure activity choice behavior of the commuters who have long activity duration, stable activity location, and strong time preference can be predicted with high accuracy.