Understanding Urban On-Street Parking Violations: A Bayesian Quantile Regression Analysis

Abstract

On-street parking violations has significant adverse effects on both traffic efficiency and safety. Existing studies are constrained by data limitations, leading to an absence of a comprehensive analytical framework for examining the impact mechanisms of parking violations across temporal dimensions. This gap impedes a deeper understanding and effective management of parking violations. To uncover the temporal patterns and determinants of parking violations in urban areas, this study analyzed 10,396 electronic records of parking violations captured by law enforcement over a two-month period. Nineteen influencing factors were selected from three key domains: land use, parking supply, and road design. Hierarchical clustering and Pearson correlation analysis were employed to examine the temporal distribution of parking violations and to refine feature selection. Given the characteristics of the dependent variables, a Bayesian quantile regression model was developed to evaluate the relationship between parking violations and influencing factors during different peak hours. The key findings of this study are as follows. First, concerning the temporal distribution of parking violations, weekday violations occur at a significantly higher frequency than those on weekends, with the highest concentration observed in the middle of the week. Clustering analysis of weekday time intervals identified three distinct patterns: peak hours (8:00–10:00 and 17:00–19:00), periods of peak-hour convergence (10:00–11:00 and 15:00–17:00), and other time periods. Second, the Bayesian quantile regression results reveal that land use, parking supply, and road design exhibit nonlinear effects on parking violations. Therefore, parking violation management strategies should be tailored to specific temporal and spatial contexts.