Procedure for Traffic Accident-Prone Area Monitoring Based on Kernel Density Estimation

To prevent traffic accidents, a variety of interventions can be implemented. Among many causes of traffic accidents, the one caused by the failure of road systems is something that should be able to be prevented by continuously monitoring the historical data of accidents. The prevention of traffic accidents via accident data analysis necessitates comprehensive procedures to ensure that continuous improvement is happening. The procedure utilizes kernel density analysis to pinpoint the accident-prone area to monitor. Accident-prone areas are classified into five categories: “Very High,” “High,” “Moderate,” “Low,” and “Very Low.” In addition, overlay mapping of “Very High” category from different years provides the prioritization of accident-prone areas to monitor. The top priority for monitoring is the intersection of the “Very High” overlay from consecutive years. The accident-prone area to be prioritized is then observed to investigate the cause of the problem and then provide recommendations to improve the road systems. The effort should become the responsibility of many relevant stakeholders. The recommendations are sent to the relevant stakeholders for further action. A demonstration of the procedure was conducted in a regency in Indonesia with a high accident rate, resulting in a successful analysis of traffic accident-prone monitoring.