Getting there first : real-time detection of real-world incidents on Twitter

Social networking and micro-blogging services such as Twitter have become a valuable source of information on current events. Widespread use of Twitter on mobile devices and personal computers enables users to share short messages on any subject in real-time, thus making it suitable for early detection of unexpected events where fast response is critical. In this paper, we present an online method for detection of real-world events in Twitter data, such as natural disasters or man-made catastrophes, by analyzing Twitter data. Our method combines different textual and frequency components that represent or approximate interesting semantic aspects of an event. We use visualization as a validation vehicle, which allows us to understand which components are relevant and what impact the parameters have on the results of our event detection algorithm.