Clustering Analysis of Traffic Accident Dataset using Canopy K Means

Due to increased urbanization, many states in India faces severe road traffic. This leads to more congestion and there by road safety. Identifying the road accident zones, causes and remedial actions needs data analytics. In this paper we are proposing a clustering method to identify road accident hotspots. The data related to traffic accidents are collected from the Kaggle website related to various attributes. After that, the processing of the data is performed to remove the invalid data and also replace the missing attributes. The number of accidents is calculated with the help of the data and Eigen values and Euclidian distance and clustering approach. Finally, this identification of road accident hotspots are given to authorities and policy makers for taking more precautions in the states which are labelled to have higher accidents.