Context Analysis of Volunteered Geographic Information from Social Media Networks to Support Disaster Management

The increasing diffusion of integrated mobile devices connected with social networks has changed the way geographic information is collected, distributed and used. Several studies have already investigated the impact of social media during crisis events. Although networks of volunteers have demonstrated the ability to curate a large amount of information reliably, this approach faces issues of sustainability and scalability. Therefore, the authors propose a semi-automatic approach to extract volunteered geographic information from social media networks, to evaluate its quality, and thereby to render it useful during any crisis event. The system presented is novel in its approach in that it focuses less on individual pieces of information, and instead uses (geographic) context to determine quality and utility. This paper presents a successful case study on forest fires, but the system architecture is adaptable to different types of crisis events.