A Data-Driven Approach for Estimating Road Utilization: Analysis and Applications

Accurate estimation of traffic flow and road utilization is crucial for urban planning and environmental sustainability. Existing methods often struggle with incomplete data and variable traffic conditions. To address these challenges, we propose a novel framework that leverages GPS trajectory data, road characteristics, and points of interest (POIs). Our framework introduces a probabilistic graphical model (PGM) to estimate traffic flow and a latent parameter road efficiency for estimating road utilization. Thereafter, we calculate a weighted utilization matrix integrating speed, volume, and road efficiency that provides a comprehensive evaluation of road network. Our evaluation of the Beijing traffic network shows a Mean Relative Error (MRE) ranging from 0.54 to 0.82 in estimating traffic flow and uncovers distinct utilization patterns across various road hierarchies. We also analyze fuel consumption and emissions in conjunction with traffic parameters. These findings reinforce the impact of road utilization on environmental factors, highlighting opportunities for targeted interventions to enhance efficiency.