KIT | KIT-Bibliothek | Impressum

A step towards real-time analysis of major disaster events based on tweets

Dittrich, André; Lucas, Christian

Abstract (englisch): The most popular micro blogging platform Twitter has been the topic of a variety of research papers related to disaster and crisis management. As an essential first step and basis for a real-time methodology to exploit Twitter for event detection, localization and ultimately semantic content analysis, a functional model to describe the amount of tweets during a day has been developed. It was derived from a corpus of messages in an exemplary area of investigation. To satisfy the different daily behavior on particular days, two types of days are distinguished in this paper. Moreover, keyword-adjusted data is used to point out the potential of semantic tweet analysis in following steps. The consideration of spatial event descriptions in relevant tweets could significantly improve and accelerate the perception of a disaster. The results from the conducted tests demonstrate the capability of the functional model to detect events with significant social impact in Twitter data.

Zugehörige Institution(en) am KIT Institut für Photogrammetrie und Fernerkundung (IPF)
Publikationstyp Proceedingsbeitrag
Jahr 2013
Sprache Englisch
Identifikator ISBN: 978-3-923704-80-4
KITopen ID: 1000058061
Erschienen in 10th International Conference on Information Systems for Crisis Response and Management (ISCRAM'13), Baden-Baden, 12.-15.05.2013. Ed.: T. Comes
Seiten 868-874
Schlagworte event detection, disaster, social sensor, Twitter
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft KITopen Landing Page