GIS-based crash hotspot identification: a comparison among mapping clusters and spatial analysis techniques.

Area of Science:

• Transportation Science
• Spatial Analysis
• Road Safety Engineering

Background:

• Understanding traffic crash hotspots is vital for identifying root causes and informing safety interventions.
• Effective risk assessment and countermeasure development rely on accurate hotspot identification.
• Locating areas with a high potential for traffic crashes presents a complex challenge.

Purpose of the Study:

• To compare the effectiveness of five different hotspot identification techniques for traffic crashes.
• To evaluate these techniques using multiple performance metrics: Predictive Accuracy Index (PAI), Recapture Rate Index (RRI), and hit rate.
• To determine the most accurate and reliable method for traffic crash hotspot analysis.

Main Methods:

• Comparative analysis of five hotspot identification techniques: Average Nearest Neighbor, Getis-Ord Gi*, Global Moran's I, kernel density estimation (KDE), and mean centre.
• Performance evaluation using Predictive Accuracy Index (PAI) and Recapture Rate Index (RRI).
• Statistical assessment of accuracy and reliability across different spatial analysis methods.

Main Results:

• Global Moran's I demonstrated the highest accuracy, yielding the highest PAI values (1.61 and 1.76).
• The Getis-Ord Gi* method showed the highest reliability (RRI = 1.121) but had the second-lowest accuracy (PAI = 0.83 and 0.74).
• Global Moran's I achieved a high reliability score (RRI = 1.003), ranking third overall.

Conclusions:

• Global Moran's I is identified as a superior method for traffic crash hotspot identification due to its high accuracy.
• The study suggests that Global Moran's I offers a reliable and precise approach for locating areas prone to traffic incidents.
• The findings provide valuable insights for traffic safety researchers and practitioners in selecting appropriate hotspot analysis tools.