Step by Step for Road Safety

A Study on Accident Black Spots

Road safety is one of the leading problems of modern societies, and every year thousands of people are killed or injured in traffic accidents. In place of traditional approaches to road safety, the concept of "Vision Zero" — known as the systems approach to road safety — has emerged. This approach aims to eliminate traffic fatalities entirely. The Vision Zero approach requires focusing on reducing the likelihood of accidents by examining their causes. For this reason, the analysis of accident data is important for developing solution proposals toward this goal. These analyses are important for improving safety culture, identifying hazardous conditions and behaviors, and enabling predictability and proactive measures.

The analysis of traffic accidents provides valuable information for reducing accident frequency, minimizing injuries and losses, and creating a safer environment overall. Reducing traffic accidents through data-driven analyses and appropriate measures aims to ensure the safety of drivers, pedestrians and other road users.

Analyses of accident data are important tools used to make critical road-safety decisions. Data analysis and statistical methods have become a powerful tool for understanding traffic accidents, identifying their causes and taking effective measures. Accident black-spot analysis enables the identification of the areas or points where traffic accidents are concentrated. Identifying accident black spots reveals priority areas in terms of traffic management and infrastructure improvements. In this way, the aim is to prevent accidents by taking safety measures in these areas.

Descriptive Analysis of Accident Data

Descriptive analysis of traffic accidents is a method used to understand the distribution and pattern of accidents over time. This analysis aims to understand the factors affecting traffic by determining in which years, months (Figure 1), seasons, days or hours (Figure 2) accidents are concentrated. For example, it can be observed that more accidents may occur during certain time periods, or that weather conditions in certain seasons may increase accidents. During hours when accidents are frequent, traffic flow is more likely to be affected. For example, an increase in accidents may be seen during evening rush hours or holiday periods. The most frequent accident types can be identified and location, time and accident type can be matched.

This information can be used to determine the hours during which traffic controls are increased or alternative routes are recommended. It can provide important information for planning and implementing road-safety measures. For example, traffic control can be increased during hours when accidents are frequent, or special warnings can be issued to drivers in risky seasons.

Spatial Examination of Traffic Accidents and Identification of Accident Black Spots

Spatial analysis of traffic accidents is a method that enables the identification of risky areas by examining the areas where traffic accidents occur. This analysis involves mapping the points where accidents occur based on their geographic locations. In areas where accidents are concentrated, detailed analyses can be performed on factors such as driver errors, intersections and road conditions. Different methods can be used depending on the purpose of the study and the scale of the study area: Getis-Ord Gi*, Anselin Local Moran's I, Nearest Neighbor Hierarchical Clustering and the Empirical Bayes Method.

The hot-spot analysis results based on the Getis-Ord G* method are shown in Figure 4, and the results based on the Anselin Local Moran's I method are shown in Figure 5. According to these results, the high-density city center and its surroundings (red points) were identified as accident hot spots with a 90-99% confidence level. An accident hot spot indicates that there are more accidents than expected, i.e. high-density accident areas. These are areas with interrelated or clustered accidents. Areas that can be considered the city periphery were identified as accident cold spots with a 90-99% confidence level. An accident cold spot indicates fewer accidents than expected, i.e. low-density accident areas.

While these methods focus on regional clustering when detecting accident hot spots, accident black spots in the city center were also identified using the "Nearest Neighbor Hierarchical Clustering" method to obtain more detailed results at the intersection and road-segment level. Within the scope of this study, accident black spots in the Nearest Neighbor Hierarchical Clustering method were calculated using threshold distances of 1,000 m, 500 m, 200 m, 100 m and 75 m and minimum parameters of 5, 10, 20 and 50 points (Figure 6, Figure 7, Figure 8, Figure 9, Figure 10).

The Empirical Bayes Method was applied to road segments and intersection points for all years, and accident prediction data were obtained at each point for the five-year period between 2017 and 2021. The resulting accident predictions were compared with the recorded data collected from accident points to identify accident black spots for road segments and intersections. The results obtained by the Empirical Bayes method applied at a selected intersection are given in Table 1. In the accident predictions made using the intersection's existing traffic volume and characteristics with the Empirical Bayes Method, a total of 37 accidents were expected over the five-year period, while the observed number of accidents was 58.

Results that can be obtained by identifying accident black spots:

• Root causes of traffic accidents: Identifying accident black spots makes it possible to understand the root causes of accidents that frequently recur in a specific area or road section.
• Safety measures that need to be improved: Identifying black spots highlights the safety gaps in that area or road section. For example, it may indicate that measures such as adding barriers or warning signs, making road adjustments, or changing traffic signals need to be taken.
• Solution proposals to improve traffic flow: Understanding how a particular black spot affects traffic flow makes it possible to develop solution proposals to reduce congestion in the area. Various measures such as road widening, creating alternative routes, or reconfiguring traffic signalization can be evaluated at this point.
• Analysis of user behavior: Examining accidents at black spots offers the opportunity to understand and evaluate driver behavior. For example, behind accidents that frequently occur in a particular area there are often certain behavior patterns such as speeding, careless driving, or driving under the influence / in poor condition.
• Strategies to prevent future accidents: Identifying black spots helps develop strategies to prevent future accidents. Data analysis and examination of traffic patterns can guide the planning and implementation of preventive measures.