Short-distance commuting mode selection mechanism considering spatial heterogeneity

Short-distance traveling is a vital component of students' daily commuting routines. Guiding the mode choices of short-distance school commuters is crucial for fostering healthy student development and alleviating road traffic congestion during peak hours. Using Nanjing city as a case study, this research used Bayesian classification to identify short-distance travel distance threshold for student groups and found that the short-distance travel threshold for students is 2.7 km in Nanjing. Subsequently, a multinomial logit (MNL) model and a geographically weighted multinomial logit (GWMNL) model were constructed to explore the effects of personal attributes, travel characteristics, family characteristics, and built environments on the mode choices of schooling for groups of short-distance students, considering spatial heterogeneity. The results show that the GWMNL model exhibits superior fitting and explanatory power compared to the MNL model, indicating significant differences in students' commuter travel mode choice behaviors depending on their residential locations. For instance, when compared to younger students residing closer to their schools, senior students in urban and suburban areas, despite having shorter travel distances, are more likely commute by cars. Conversely, senior students in exurban areas, such as Liuhe and Jiangning districts, tend to prefer active travel and electric bikes for their school commute. The number of cars shows a negative correlation with students' preferences for active travel, electric bikes and public transport when commuting to school. This negative impact progressively intensifies from the city center towards the outskirts. Additionally, the distance to the nearest bus stop has a positive effect on students' choice of public transport in the urban and suburban areas. However, this positive impact turns negative in some more remote exurban areas, particularly in the northern and eastern parts of Nanjing. The findings of this study provide valuable insights for the development of active travel (walking and cycling) guidance strategies for short-distance commuting student groups of students in different regions.