Modeling spatial dimensions of parcel delivery demand and its determinants

This study aims to investigate the spatial dimensions of parcel delivery demand and its key determinants. Based on the parcel delivery volume data in South Korea, we regressed per capita parcel delivery volume on socioeconomic, land use, housing, and industry-related variables of city-level districts. When covering the entire country, Spatial Durbin Error Model worked best in modelling the spatial dependence in the disturbance and the spatial spillover effect of the explanatory variables in the neighbouring districts. When performing separate regression analyses for urban and non-urban areas, we identified a different relationship between parcel demand and its determinants in two subgroups. Given the rapid development of e-commerce, establishing logistics infrastructure for supporting home delivery services, particularly in urban areas, has become challenging. This study can contribute to improved prediction of parcel demand that varies over space, suggesting region-specific policies for establishing local parcel delivery infrastructure such as sub-terminals.