Twitter Data Mining for Situational Awareness

Abstract

The most recent catastrophic events, from the 2010 Haiti earthquake to the devastating 2013 Colorado floods, have shown a strong adoption of social media platforms by ordinary people. The data and meta-data produced by the users during and after the extra-ordinary situations could have enormous potentialities if integrated with the traditional systems for emergency management and used for hyperlocal situational awareness. The great majority of the current literature is focused on Twitter for several reasons strictly linked to the architectures and practices of use of the platform itself. It is possible to classify the existing systems based on the analysis of Twitter data at least in three different categories: 1) semantic systems; 2) meta-data systems; and 3) smart self-learning systems. In this contribution, a review of the most significant and important tools used to analyze Twitter data will be presented and an innovative and smart solution will be proposed for future development.