Tweeting About Floods of Messinia (Greece, September 2016) - Towards a Credible Methodology for Disaster Management Purposes

Abstract

As repeatedly stated by many researchers, the phenomenon of volunteered geo-graphic information (VGI) is an important new technological tool that can contribute to disaster management (DM). Specifically regarding floods, despite that there is very interesting published research work that assesses VGI and flood event management in various areas of the world, related research in Greece, is still limited. Moreover, the necessity of creating meaningful maps and graphs is still a significant challenge glob-ally, along with the general need to reduce the time needed for data processing. Cur-rent research presents steps towards those issues, by advancing various innovative initial methods and techniques that have been presented previously. The floods of Kalamata, Messinia (Greece, September 2016) were selected for case study, while as a VGI source, a corpus of 111000 tweets, published within 120 hours since the flood event occurrence, was used. As far as the methodology is concerned, the main steps, included: a. the classification of the tweets in categories, the main of which were: i) flood tracking, ii) tracking of consequences, iii) simple identification of rain and iv) irony/emotions. The next step was related to geo-referencing the classified infor-mation, by using an R script, while the final steps were linked to generating maps and graphs, from subsets of the processed data. The outcome of this research is based on the analysis of about 45% of the total dataset.