Towards Next-Generation Software Infrastructure for Crisis Informatics Research

Abstract

Crisis Informatics is a multidisciplinary research area that examines the socio-technical relationships among people, information, and technology during mass emergency events. One area of crisis informatics examines the on-line behaviors of members of the public making use of social media during a crisis event to make sense of it, to report on it, and, in some cases, to coordinate a response to it either locally or from afar. In order to study those behaviors, this social media data has to be systematically captured and stored in a scalable and reliable way for later analysis. Project EPIC is a large U.S. National Science Foundation funded project that has been performing crisis informatics research since Fall 2009 and has been designing and developing a reliable and robust software infrastructure for the storage and analysis of large crisis informatics data sets. Prof. Ken Anderson has led the research and development in this software engineering effort and will discuss the challenges (both technical and social) that Project EPIC faced in developing its software infrastructure, known as EPIC Collect and EPIC Analyze. EPIC Collect has been in 24/7 operation in various forms since Spring 2010 and has collected terabytes of social media data across hundreds of mass emergency events since that time. EPIC Analyze is a data analysis platform for large social media data sets that provides efficient browsing, filtering, and collaborative annotation services. Prof. Anderson will discuss these systems and also present the challenges of collecting and analyzing social media data (with an emphasis on Twitter data) at scale. Project EPIC has designed and evaluated software architectural styles that can be adopted by other research groups to help develop their own capacity to work in this space. Prof. Anderson will conclude the talk with a vision for future work in this area: What's next for crisis informatics software infrastructure?