Crowdsourced Traffic Accident Reporting for Hot Spot Analysis and Emergency Response

Abstract

Traffic accidents are a major cause of death worldwide. Rapid response by emergency medical personnel is essential to reduce fatalities. However, preventing accidents is equally important. Although many accidents are caused by human error, infrastructure design and conditions also play a role. Thus, government agencies concerned with both public health and transportation planning want to identify locations where accidents frequently occur. This activity, often called ``hot spot analysis{''}, requires detailed spatio-temporal information about accident events. Such information can be difficult to obtain. Relevant data are often locked in paper reports or text documents rather than stored in geospatially-enabled data bases. Even when resources are available to convert data to a digital form that can be analyzed and mapped, there will be a significant time lag between accident occurrence and analysis. In this paper, we report on a system that addresses these issues. Our EMTAlert system provides emergency medical personnel with near-real-time notification about accident events while also gathering and storing geo-tagged accident data for analysis and modeling. EMTAlert consists of a free mobile application that anyone can use to report traffic accidents, a geospatial data base to retain accident data, and a web application for querying and displaying accident data on a map. Accident reports created by EMTAlert can be immediately viewed by monitoring personnel and used to dispatch an ambulance if necessary. Meanwhile, accident details are retained to serve as the basis for reporting, time series displays, hot spot analysis, and simulation.