The Car that Hit The Burning House: Understanding Small Scale Incident Related Information in Microblogs

Abstract

Microblogs are increasingly gaining attention as an important information source in emergency management. In this case, state-of-the-art has shown that many valuable situational information is shared by citizens and official sources. However, current approaches focus on information shared during large scale incidents, with high amount of publicly available information. In contrast, in this paper, we conduct two studies on every day small scale incidents. First, we propose the first machine learning algorithm to detect three different types of small scale incidents with a precision of 82.2% and 82% recall. Second, we manually classify users contributing situational information about small scale incidents and show that a variety of individual users publish incident related information. Furthermore, we show that those users are reporting faster than official sources