Abstract

Every year, India loses nearly 1.5 million citizens due to road accidents. In this work, we try to implement a two pronged approach to reduce the number road accidents in India. Firstly, we create a Machine learning model to predict road accidents subject to different road categories in India, using data from publicly available government dataset. Parallelly, we have a web application that we can use to decentralize the mapping of accident hot-spots via OpenStreetMaps. The hot spot markers vary in their intensity depending upon the accidents occurring in that locality. We also categorize the accidents into National highways, State highways and Other roads for more accurate measures so that various institutions can use them to make policies to reduce the mortality of road accidents in India. The results derived from the model are also presented that demonstrates the high efficiency of the model.