Spatial analysis and predictive modeling framework of truck parking and idling impacts on environmental justice communities

Abstract

This study introduces a comprehensive modeling framework to analyze truck idling and parking activities, illustrated through a case study in environmental justice communities in Kern County, California. It includes 1) exploratory spatial and cluster analysis to identify hotspots of those truck activities and their influencing factors, and 2) advanced predictive models, particularly the Cross-Validated Random Forests model, to predict and investigate critical factors influencing truck idling time, parking search time, and inferred truck parking demand. The results reveal that the percentage of heavy-duty trucks and the specific land use influence truck idling time. For parking search time, key predictors include distance to major roads and employment in certain industries. The inferred truck parking demand model underscores the impact of commercial land use areas, proximity to major roads, and socioeconomic factors. These findings enable the identification of hotspots for truck idling and parking searches, facilitating targeted interventions such as optimizing land use planning, improving infrastructure around major roads, and enhancing parking facilities in commercial zones. Integrating spatial, socioeconomic, and GPS aggregate data, the methodology provides a scalable framework applicable to other regions facing similar challenges through data-driven planning and policy initiatives.