Towards better parking provision: Insights from parking lot utilization analysis of Hangzhou, China

Abstract

Parking provision is a key concern in parking planning and management. Policymakers and parking practitioners often over-supply parking due to fears of shortages, leading to inefficient and unstrategic allocation. There is limited understanding of how many existing parking spaces are used and what factors affect their use. Most past research has focused on residential or on-street parking, leaving travel-related parking, especially at the parking lot level, understudied. This study analyzes data from 2592 public parking lots with over 700,000 spaces in Hangzhou, China, to assess parking utilization and explore influencing factors. Using advanced machine learning techniques, including XGBoost and SHAP values, the study investigates nonlinear relationships between parking utilization and factors such as parking lot attributes, land use, urban accessibility, and trip factors. Results reveal a significant oversupply of parking in Hangzhou, with average parking lot utilization rates of 0.57 on workdays and 0.48 on non-workdays. 56 % of parking spaces remain unused during workdays, and 78.2 % of parking lots are oversupplied (utilization rate below 0.85). SHAP analysis shows that small-scale lots have higher utilization rates compared to large-scale ones. Heterogeneous land use leads to significant differences in utilization; facility aggregation increases utilization, while a high land use mix and good public transportation accessibility reduce it. Given the substantial amount of unused parking observed, this study emphasizes the need for better parking provision policies in cities. The findings offer policy implications for improving parking provision and promoting a sustainable urban parking market.