Graphical Style Sheets: Reusable Representations of Domain-Specific Information Graphics

Ramon Felciano, Stanford Section on Medical Informatics

Seminar on People, Computers, and Design
Stanford University April 10, 1998


Over time, practitioners in particular scientific or technical domains may custom-tailor data graphics to display certain types of information. Biomedicine in particular has many specialized graphics used to display genetic, molecular, and physiologic data. As these statistical and informational graphics become increasingly specialized, their value to biomedical practitioners increases, while their applicability to data from other fields diminishes. If these specialized graphics, which we term "Domain Graphics", become widely used and accepted by practitioners in the field, they become part of the lingua franca of that field. Domain graphics differ from generic visual formalisms in their use of domain-specific layout rules, information symbols, and other drawing conventions in the display of domain data. As they become more specialized and specifically useful, they become difficult to create because of the relative lack of specialized drawing and visualization tools.

To support the use of domain graphics in interactive information visualizations and specialized user interfaces, we propose to model domain-specific graphic design conventions as declarative, knowledge-based graphical style sheets. Graphical Style Sheets (GSS) define the layout and drawing conventions shared by members of a particular family of data graphics. We have developed a conceptual that describes how a visualization system can use graphical style sheets to determine how to represent domain data elements as graphical elements positioned on a page (the essence of information visualization). The framework includes (1) an object-oriented representations of data, (2) a numeric and logical metrics that measure and constrain the characteristics of graphical elements (e.g. position, color) and (3) a declaration of mappings between domain concepts and these graphical objects and constraints. We have developed a prototype constraint-based visualization system (PALLADIO) and design representation language (P-SPEAK) to evaluate this framework.

This is joint work done with Russ Altman at the Stanford Section on Medical Informatics.


Ramon M. Felciano is a Ph.D. candidate in Medical Informatics at the Stanford Section on Medical Informatics. He holds bachelors degrees from Stanford in Computer Science and in English and French Literatures. Mr.Felciano¹s research interests include innovative methods for designing intelligent user interfaces, interactive information graphics, and Internet technologies. Prior to joining the Section on Medical Informatics, Mr.Felciano co-founded SUMMIT, the Stanford University Medical Media and Information Technologies lab, where he held the position of Associate Director for 4 years. Mr. Felciano is also the founder of Digital Alchemy, a design and consulting firm based in Palo Alto, California.


Titles and abstracts for all years are available by year and by speaker.

For more information about HCI at Stanford see

Overview Degrees Courses Research Faculty FAQ