Identifying the critical features influencing warehouse rental prices and their nonlinear associations: A spatial machine learning approach

Abstract

Warehouses play a crucial role in freight transportation, and their pricing strategies affect warehouse location choices and associated environmental impacts. Although most firms rent storage spaces, limited studies have examined warehouse rental prices (WRP). Furthermore, most studies assume a pre-defined relationship between WRP and its correlates. This study applies spatial machine learning models to warehouse rental data in Shanghai to examine their nonlinear associations. The results show that the primary factors influencing WRP include spatial dependence among warehouses, location and neighborhood attributes, and the floor level of warehouse spaces, whereas lease and service-related factors contribute minimally. Moreover, spatial dependence leads to segmented markets, with high-rent warehouses clustering in the central urban area and around logistics parks and transportation terminals outside the central area. Additionally, most primary correlates exhibit irregular nonlinear relationships with WRP, which shed light on warehouse pricing mechanisms and provide guidance for location choices.