Predictive Model of Pedestrian Crashes Using Markov Chains in the City of Badajoz

Driving a vehicle, whether motorized or not, is a risky activity that can lead to a traffic accident and directly or indirectly affect all road users. In particular, road crashes involving pedestrians have caused the highest number of deaths and serious injuries in recent years. In order to prevent and reduce the occurrence of these types of traffic accidents and to optimize the use of the available resources of the administrations in charge of road safety, an updatable predictive model using Markov chains is proposed in this work. Markov chains are used in fields as diverse as hospital management or electronic engineering, but their application in the field of road safety is considered innovative. They are prediction and decision techniques that allow the estimation of the state of a given system by simulating its stochastic risk level. To carry out this study, the available information on traffic accidents involving pedestrians in the database of the Local Police of Badajoz (a medium-sized city in the southwest of Spain) in the period 2016 to 2023 were analyzed. These data were used to train a predictive model that was subsequently used to estimate the probability of occurrence of a traffic crash involving pedestrians in different areas of this city, information that could be used by the authorities to focus their efforts in those areas with the highest probability of a road crash occurring. This model can improve the identification of high-risk locations, and urban planners can optimize decision making in designing appropriate preventive measures and increase efficiency to reduce pedestrian crashes.