Cloud based spatial visualization with statistical approach for road accidents

Abstract

Road accidents have emerged as a major public health problem globally. The scope of road accident data usage is limited to serve statistical purposes. Therefore it is necessary to identify high-frequency accident locations by performing risk evaluation using Geographic Information System (GIS) and statistical analysis to enhance road safety. GIS is a powerful spatial analytics system for accident pattern analysis. Our aim is to provide a cloud based GIS application for Roorkee City that visualizes the analysis of hotspots in raster as well as vector format using kernel density estimation, buffer analysis and nearest neighborhood analysis for accident and hospital locations over the cloud to achieve shortest path from accident locations to nearest hospitals and police stations. The accident locations have been classified into five zones i.e. very high accident frequency zone, accident frequency zone, moderate accident frequency zone, less accident frequency zone and very less accident frequency zone using heat map analysis with the help of kernel density estimation. Quantum-GIS have been applied for heatmap analysis, buffer analysis and nearest neighborhood analysis over the cloud. Periodically the database will get updates and reduce accident severity. Any user can get any information about road accident attributes with coordinates, date and time over the cloud and will be of benefit for decision-making processes.