Using 9-1-1 call data and the space–time permutation scan statistic for emergency event detection

Abstract During medium to large-scale emergency events such as fires or earthquakes, emergency response teams must be deployed to the location of the event in a timely fashion. Information gathered from 9-1-1 call makers is useful for this, but is not always accurate. In this paper, we show how the space–time permutation scan statistic can be used to detect the presence and location of such events by monitoring the overall spatiotemporal pattern of 9-1-1 emergency calls instead. In an analysis of collected 9-1-1 call data, top detected clusters were found to correlate with emergency events as reported on the news, showing the algorithm's usefulness for automatically estimating their location and temporal extent. We show how the detection procedure works in cases where the emergency event generates a small but statistically significant increase in the number of 9-1-1 calls, as well as cases where events must be detected against a large background activity of 9-1-1 calls. We discuss the possible adoption of the proposed algorithm within the next generation digital government framework.