Beyond Seeing Patterns and Relationships

Steven Roth
Carnegie Mellon University
Pittsburgh, PA
USA

Information visualization cannot be limited to supporting the process of seeing patterns and relationships in data. More broadly, people working in information-intensive domains also need to manipulate information to search, filter, control level of detail, reorganize, change, and derive new information. Furthermore, it is rare that people merely analyze information - they typically need to communicate and collaborate with others about it. Ultimately, visualizations of information must enable people to control the world - both electronic and otherwise. Visualization must serve as application user interfaces to control systems. The latter implies the need for new user interface environments - that is, new workspaces - within which visualizations can serve these diverse purposes.

We are developing an approach to such an environment called Visage, with which we are exploring the seamless integration of these capabilities. We have used Visage to create both customized, special-purpose "information appliances" designed for users to achieve a single analysis task. We have also created general-purpose "tools" for navigating, filtering, aggregating, and performing other basic operations on information throughout all applications constructed in the Visage environment. In contrast to appliances, tools are composed and applied by knowledgeable users to perform novel tasks not anticipated by developers. This talk will provide an overview of the Visage environment and discussion of some of the interesting design problems arising in the coordination of these two different types of visualization interface.


Program
Session 2: Visualization of Database Information
Chris Ahlberg (unavailable)
Beyond Seeing Patterns and Relationships by Steven Roth
Visualization Techniques for Mining Large Databases: A Comparison by Daniel Keim

Table of Contents