Understanding fly-tipping in urban areas: A social-economic-spatial combinatorial approach enabled by geographically weighted random forest

This study addresses the ongoing challenge of illegal dumping (‘fly-tipping’) in urban environments, specifically focusing on Hong Kong. Utilizing a substantial dataset of 1384 Fixed Penalty Notices issued from 2014 to 2021, the research presents a novel methodological approach that integrates Space Syntax Analysis (SSA) and Geographically Weighted Random Forest (GW-RF). The integration demonstrates robust predictive performance (R 2 = 0.809), significantly outperforming traditional models. Our findings reveal three fundamental patterns: social-economic pressures serve as consistent underlying drivers, urban spatial configuration creates distinct opportunity patterns, and waste disposal infrastructure acts as a local modifier of dumping behavior. These insights challenge conventional assumptions about illegal dumping locations and effectiveness of uniform solutions. The study provides an evidence-based foundation for developing targeted interventions, suggesting that effective policies must consider both stable social-economic pressures and variable spatial-infrastructural conditions. This research not only deepens the understanding of fly-tipping in Hong Kong but also proposes a robust, adaptable analytical framework for other urban contexts, contributing to both urban theory and practice in environmental management. • Integration of SSA and GW-RF reveals patterns in illegal dumping behaviors. • Model achieves strong predictive performance (R 2 = 0.809) in fly-tipping analysis. • Identified key social-economic-spatial factors influencing fly-tipping. • Decoded factor significance for illegal dumping insight. • Offered a framework for targeted urban waste management.