Supporting synthesis in geovisualization

Geovisualization tools are intended to support analysts in complex task domains like crisis management, disease surveillance, and threat analysis. It is likely that analysts in these domains will use geovisualizations to develop many analytical results over time. This calls for attention to the problem of collecting, organizing, and making sense out of groups of analytical results - a stage of analysis called synthesis. The research reported here aims to characterize the process of synthesis as it is conducted by analysts working alone, and to suggest design guidelines for new tools to support synthesis in that setting. We have developed a new experimental method for observing and characterizing the process of synthesis. This approach has participants work with a collection of physical data artifacts on a paper-covered workspace to devise hypotheses under the guise of a disease outbreak scenario. From experiment video recordings we identified and coded actions that participants undertook to complete the synthesis task. In this article we report results from synthesis experiments with analysts from Pacific Northwest National Laboratory and experts from The Pennsylvania State University. Experiment results are then distilled into a design framework that can be used to shape the development of geovisual synthesis tools.