Data visualization plays a fundamental role in helping users solve complex, information-intensive problems in scientific, engineering and business applications. This thesis introduces an original framework for the use of queries to create and filter structural data visualizations (a term we introduce to refer to the diagrammatic display of the relationships of structured data). Hygraphs, a new visual formalism, provides a precise characterization of the diagrammatic visualizations considered. This simple formalism is a convenient abstraction for both graph-based and forin-based (or tabular) presentations.

We present theoretical and practical contributions that exploit the synergism between the established field of database query languages and the emerging area of visualization. On the database theory side, we introduce the concept of filter queries and provide a formal definition and expressive power characterization of filtering programs. We then extend the visual query language GraphLog to make use of hygraph patterns for both defining derived data and expressing filter queries. The framework presented here supports the creation of structural data visualizations by using hygraph patterns to both define new relationships in the data to be visualized and filter the existing data to display only information that is relevant to the user.

On the practical side, we describe the Hy⁺ Hygraph Visualization System. Hy⁺ embodies a significant amount of the functionality that can be developed within the formal framework described in this thesis. We discuss the application of Hy⁺ to software engineering and network management to demonstrate the possibilities opened up by creating and filtering structural data visualizations using hygraph patterns.