Abstract
Social networks are commonly used by citizens as a communication channel for sharing their messages about a crisis situation and by emergency operation centers as a source of information for improving their situation awareness. However, to utilize this source of information, emergency operators and decision makers have to deal with large and unstructured data, the content, reliability, quality, and relevance of which may vary greatly. In this paper, to address this challenge, we propose a visual analytics solution that filters and visualizes relevant information extracted from Twitter. The tool offers multiple visualizations to provide emergency operators with different points of view for exploring the data in order to gain a better understanding of the situation and take informed courses of action. We analyzed the scope of the problem through an exploratory study in which 20 practitioners answered questions about the integration of social networks in the emergency management process. This study inspired the design of a visualization tool, which was evaluated in a controlled experiment to assess its effectiveness for exploring spatial and temporal data. During the experiment, we asked 12 participants to perform 5 tasks related to data exploration and fill a questionnaire about their experience using the tool. One of the most interesting results obtained from the evaluation concerns the effectiveness of combining several visualization techniques to support different strategies for solving a problem and making decisions.