Characterization of road accidents and governmental actions for the reduction of deaths and injuries in transit

Abstract

This study aims to identify the main contributing factors of accidents and to report on government actions to reduce road deaths and injuries. The characterization of traffic accidents recorded in the Federal Highway Police Department (PRF) database is performed using numerical and statistical methods, with data sample between 2014 and 2016 and considering the federal highway BR-101, which had the highest volume of traffic recorded in the period, according to data provided by the National Department of Transport Infrastructure (DNIT). In order to assist in the identification of the contributing factors of road accidents, it was made use of the knowledge discovery process in database, with the data mining technique and the Apriori algorithm. There was a reduction in the number of accidents in 2016 compared to 2014; possibly related to the intensification of PRF road inspection operations and recent educational campaigns among users of the transit system. Most of the identified causes of accidents were related to the behavior of the person driving the vehicle (incompatible speed, lack of attention, lack of safe distance, signaling disobedience, sleeping, alcohol intake and undue overrun).