A before-after full bayes multivariate intervention model to estimate the safety effectiveness of red light cameras.

OBJECTIVE: Intersection-related crashes account for approximately 40% of all crashes and tend to be more severe. Red-light running (RLR) crashes are most severe as almost half of these crashes result in injuries and fatalities. To reduce RLR crashes, agencies have been deploying red light cameras (RLCs). The main objective of this study was to evaluate the safety effectiveness of RLCs in the City of Miami Beach, Florida. METHOD: The full Bayes (FB) approach was conducted based on five treatment intersections with six RLCs and 14 comparison intersections without RLCs. The analysis focused on target crash types, including rear-end, sideswipe, and angle/left-turn/right-turn crashes, and crash severity. RESULTS: The FB analysis indicated a significant sudden drop in all types of target crashes immediately after the installation of RLCs. Compared to the before-period, the after-period experienced: fewer angle/left-turn/right-turn crashes, fewer sideswipe crashes, and more rear-end crashes. The sideswipe and angle/left-turn/right-turn crashes dropped immediately after the installation of RLCs and then continued to increase, but they were still lower than the before- period. The rear-end crashes dropped immediately after the installation of RLCs and then continued to increase, but they increased at a steeper rate. Major and minor approaches AADT, higher speed limit, longer amber time, length of pedestrian crosswalk, and number of driveways within the intersection influence area increased the frequency of total target, PDO, and FI crashes. Intersections with all-red interval more than two seconds, major approach with more than two through lanes, and minor approach with more than one through lane, on the contrary, resulted in a fewer number of the total target, PDO, and FI crashes. The treatment indicator showed that treatment intersections experienced fewer FI, angle/left-turn/right-turn, and sideswipe crashes and more total, PDO, and rear-end crashes compared to the non-treatment intersections. CONCLUSION: This study provides reliable estimates of the safety effectiveness of RLCs since it accounts for uncertainties in the data, regression-to-the-mean, and spillover effects.

Estimation of bicycle crash modification factors (CMFs) on urban facilities using zero inflated negative binomial models.

The objective of this study was to develop crash modification factors (CMFs) for bicycle crashes for different roadway segment and intersection facility types in urban areas. The study used four years (2011-2014) of crash data from Florida to quantify the safety impacts of roadway and traffic characteristics, bicycle infrastructure, and bicycle activity data on bicycle crashes. A cross-sectional analysis using Generalized Linear Model (GLM) approach with Zero Inflated Negative Binomial (ZINB) distribution was adopted to develop the relevant regression models in this study. Lane width, speed limit, and grass in the median were observed to have positive impacts on reducing bicycle crashes. On the contrary, presence of sidewalk and sidewalk barrier were found to increase the bicycle crash probabilities. Increased bicycle activity was found to reduce the bicycle crash probabilities on segments, while increased bicycle activity resulted in higher bicycle crash probabilities at intersections. Bus stops were found to increase the bicycle crash probabilities at intersections, whereas, protected signal control had a positive impact on bicycle safety. This research provides a greater insight into how various characteristics affect bicycle safety, a topic that is seldom considered by researchers and practitioners.

Data mining in road crash analysis: the context of developing countries.

The recent advancements in the field of data mining have made vast progress in extracting new information and hidden patterns from large datasets which are often overlooked by the traditional statistical approaches. These methods focus on searching for new and interesting hypothesis which were previously unobserved. Road safety researchers working with the crash data from developed world have seen encouraging success in obtaining new insight into crash mechanism through data mining. An attempt was made in this study to apply these advance methods and evaluate their performance in manifesting crash causes for Bangladesh. The study applies hierarchical clustering to identify hazardous clusters, random forest to find important variables explaining each of these clusters, and classification and regression trees to unveil their respective crash mechanisms for the road crash data of Bangladesh. The results identified several new interesting relationships and acknowledged issues related to quality of data.