Market Basket Analysis of Chicago Road Accidents

Abstract

Over the years, data mining techniques have grown increasingly popular. Machine learning has been implemented as means of discovering new information in a wide variety of applications, one of which is the analysis of road accidents and crash data. While many techniques have been applied to this problem; Market Basket Analysis can provide practical insight for road safety and accident prevention efforts through its discovery of interesting accident patterns. In this paper, association rules are investigated with regards to the nature of accidents, their causes, number of cars involved, and crash characteristics. By the results obtained, recommendations are thus offered to help reduce certain accident types such as hit and runs or rear accidents, with emphasis on the timeframe (weekday vs weekend) and traffic type (intersection-related). On another note, a significant decrease in accident rates following the COVID-19 lockdown is also illustrated, and noticeable changes in accident patterns due to the situation are briefly explored.