Abstract
Road safety focused on railway crossings has historically been very weak in Costa Rica. This has generated important consequences in the efficiency and losses of time, for both the passenger of the train system, and for the users of the road, due to the blockages and delays generated by the occurrence of crashes in these locations. Additionally, a greater number of crashes also implies a greater occurrence of major injuries, so it is necessary to reduce the number of crashes at railway crossings by applying infrastructure improvements, in addition to other approaches. This research applies Bayesian statistics and spatial statistics simultaneously to improve the predictions in the total number of crashes at each railway crossing and uses this input to classify the crossings according to their excess of crashes. The crossings with greater promise of improvement are identified and therefore their improvements should be prioritized in related budgets. The results show that the crossing in the Mercado de Mayoreo in La Sabana, the crossing of the Río Pirro in Heredia, among others, present conditions to prioritize their investment.