The Geography of Conflicts: A Comparative Assessment of Web Data.

The lack of trust is one of the most frequently addressed research gaps that hinders the practical deployment of social media information into decision-making process of disaster management. Furthermore, literature points out the lack of transferability of the developed methods to other countries and languages, the lack of long-term analyses, and the focus on specific online platforms. In order to bridge the gap between research and practice, this paper highlights in addition the need for data fusion, which allows web and social media data to be combined with commonly used information structures. Using the current Ukraine crisis as an example, this initial paper focuses on the fusion of social media, online news and humanitarian datasets. The goal is to evaluate the reliability of social media information. By using an aggregation methodology that structures the data based on spatio-temporal and semantic similarity, a binary reliability indicator is derived as a first prototype, which will serve as a basis for further developments. In addition, information quality indicators are presented that allow to evaluate the context created by the fusion and the results obtained from it. Preliminary results reveal the beneficial effect of the obtained context and reliability indicator. Furthermore, the results show that social media make a significant contribution to the dissemination of unique information, both in terms of spatio-temporal and semantic aspects.