Sensemaking across Text Documents: Human-Centered, Visual Exploration with Jigsaw

In this article we describe our research on providing techniques and tools for a particular type of sensemaking — finding embedded narratives or plots across large collections of relatively brief text documents. Our approach is decidedly human-centered and we have created the Jigsaw system as an embodiment of these principles. Jigsaw provides multiple visualizations depicting the documents, the entities within them, and connections between the two. Analysts interactively explore the entities and documents as a way to guide document selection for reading, which ultimately is necessary to synthesize the embedded plot. This article describes the challenges in this particular style of sensemaking as well as the ways in which Jigsaw is attempting to overcome those challenges.