Unveiling gender disparities in metro usage using gender-inclusive smart card data: Evidence from Nanjing, China

Abstract

Metro smart card data has enabled extensive research on usage patterns, yet most studies focus on overall station-level ridership trends, often neglecting the influence of gender on ridership patterns due to the lack of gender information. To fill this gap, this study compares the metro usage patterns between males and females utilizing smart card data with gender information from Nanjing, China. Then, multiple machine learning models are established to explore the determinants of ridership across gender groups. Results indicate that males with longer travel time and distance constitute 55.5 % of metro trips, and the time periods when males have more ridership peak at 7:00 on weekdays and at 10:00 and 17:00 on weekends. On weekdays, metro ridership in the urban area is high for both genders, with males significantly outnumbering females in ridership, while certain suburban stations show higher female ridership. In terms of spatial distribution, both males and females are mainly distributed in stations located in the main urban area on weekdays. On weekends, the proportion of female ridership slightly increases, particularly in urban areas, rising from 31.80 % to 32.14 %. Machine learning models show road density and workplace POIs exert a stronger influence on male ridership, contributing 30.81 % and 17.87 %, respectively. Residential POIs (38.24 %) and educational POIs (4.11 %) play a larger role in female ridership. Housing prices show a threshold effect on male ridership. The findings on gender disparities in metro usage underscore equity concerns and offer policy implications for equitable and efficient metro operations.