Abstract from an article in Computing Now, April 2016.

by Kenneth M. Anderson

The authors discuss their experiences designing and developing data-intensive systems in support of crisis informatics research with Project EPIC.

Abstract

The design and development of data-intensive software systems—systems that generate, collect, store, process, analyze, query, and visualize large sets of data—is fraught with significant challenges both technical and social. Project EPIC has been designing and developing data-intensive systems in support of crisis informatics research since Fall 2009. Our experience working on Project EPIC has provided insight into these challenges. In this paper, we share our experience working in this design space and describe the choices we made in tackling these challenges and their attendant trade-offs. We highlight the lack of developer support tools for data-intensive systems, the importance of multidisciplinary teams, the use of highly-iterative life cycles, the need for deep understanding of the frameworks and technologies used in data intensive systems, how simple operations transform into significant challenges at scale, and the paramount significance of data modeling in producing systems that are scalable, robust, and efficient.