Network-constrained spatio-temporal analysis and predictive modeling of road traffic accident hotspots for enhanced urban safety: a case study of Faisalabad, Pakistan.
Abstract
Objectives
Identifying the hotspots is crucial to prevent the hazardous effects of Road Traffic Accidents (RTAs). This research based on the significant current and future hotspots and temporal patterns of the RTAs of two years (2022-2023) by integrating the GIS tools with statistical analysis, investigates and reveals the future hotspots and road intersections and segments having high density of RTAs in Faisalabad city, Pakistan. The aim of this research is to identify the spatial hotspots of RTAs by analyzing the RTAs density along road network intersections, and to predict the future hotspots of the RTAs. It provides the insights for a thorough understanding of the road network's aspects leading RTAs.
Methods
To examine this, we have acquired the datasets of total 38,865 RTAs incidents, city area administrative boundaries, 4,458 km road network having 64,557 segments. Three advanced spatio-temporal analysis tools are employed. Network Kernel Density Estimation (NetKDE) to identify the dense accident segments, Repeat and Near Repeat analysis tool integrating with Predictive Zones analysis tool for predicting future hotspots are used. These tools allow to identify the highly effected road intersections and future hotspots locations and temporally variations of RTAs.
Results
This research seeks attention to the density of accidents at specific intersections and road sections, examining the part that city's road infrastructure and rush-hour traffic play in the frequencies of accidents. Important City's intersections and arterial network, such as West Canal Road, Chenab Chowk, and Clock Tower, were the primary locations of accidents. The most accidents are concentrated near commercial areas or roads that are commonly used for commuting. The X-intersection accounts for 27.4% and Crossroad intersection accounts for 24.1% of total accidents. The accuracy index of predictive model reached at the point of 1.78, that is indicating the 78% better efficiency of this model rather than randomly allocating the future hotspots. The results reveal a continuous pattern of accidents at these segments throughout both years, it also provides the precise locations, peak time and risk zones of accidents. That requires the traffic control during peak time and improvements in road infrastructure the installation of traffic signals and availability of proper walkable spaces and maintenance of footpaths.
Conclusion
This research provides a thorough examination of the spatio-temporal aspects of RTAs with futuristic approach, providing reliable and practical findings for enhancing the public safety, improving traffic management and infrastructure, for the well-being of the urban environment and living standard.