Traffic accidents are extremely common particularly in congested metropolis cities. Traffic accidents are a major cause of death globally, cutting short millions of lives per year. Therefore, a system that can predict the occurrence of traffic accidents or accident-prone areas can potentially save lives.
Since accidents don’t arise in a purely stochastic manner; their occurrence is influenced by a multitude of factors such as drivers’ physical conditions, car types, driving speed, traffic condition, road structure and weather. Studying historical accident records would help us understand the (potentially causative) relationships between these factors and road accidents, which would in turn allow us to build an accident predictor. Using a subset of these data, Cognitro built a model for predicting auto accidents in cities. This new powerful was developed to provide a car crash prediction heatmap on daily, weekly and monthly basis. The system was piloted using Abu Dhabi city data as a simulation study, reported a high level of accuracy and is currently being upgraded to include weather data and google street data.
The figures below show how week-ahead predicted accidents zones of various frequency or severity shift across the map and throughout the day. In the early morning hours, dense small accident zones (red) appear in the major city part of the downtown. Towards the afternoon, the accident zones become more spread out and with less frequency (green). In the evening, the accidents levels drop down significantly with size and frequency (light green).