Activity detection for multi-factors of ambulance demand areas: A case study in Bangkok

Abstract

One of the big challenges for the management of emergency medical service (EMS) in many urbans is a timely demand response when an emergency call occurs. To minimize patient mortality and disability, the response time of emergency medical services is critical. However, due to the continuous growth of the economic area, residential density, population density, traffic congestion, and epidemic area. Those factors are becoming complex issues for operation level planning, which need accurate and real-time demand area estimates to assign the area of responsibility for ambulance facilities while minimizing response time to emergencies and keep operating costs low. Therefore, in this article, we propose a conceptual framework for integrating multiple factors data both historical data and real-time data from social media to real-time EMS management. We propose an approach for identifying and analyzing hot spots of activity in data collections by using time-varying kernel density estimation (KDE) to convert multiple factors related to ambulance service (point locations and weight) for visualization and detection of abnormal intensities of the activity information and leads to the improvement of the EMS system. In addition, our model was applied to the case study of Bangkok EMS.