Visualizing sets to reveal relationships between constituent elements is a complex representational problem. Recent research presents several automated placement and grouping techniques to highlight connections between set elements. However, these techniques do not scale well for sets with cardinality greater than one hundred elements. We present OnSet, an interactive, scalable visualization technique for representing large-scale binary set data. The visualization technique defines a single, combined domain of elements for all sets, and models each set by the elements that it both contains and does not contain. OnSet employs direct manipulation interaction and visual highlighting to support the easy identification of commonalities and differences as well as membership patterns across different sets of elements. We present case studies to illustrate how the technique can be successfully applied across different domains such as bio-chemical metabolomics and task & event schedule.
At the Information Interfaces Lab, computing technologies are developed that help people take advantage of information to enrich their lives. The lab group develops ways to help people understand information via user interface design, information visualization, peripheral awareness techniques and embodied agents. The goal is to help people make better judgments by learning from all the information available to them.