Hy+: A Hypergraph-based Query and Visualization System

This paper provides abrief overview of Hy+, ahygraphbased query and visualization system, which is presented in the accompanying video. Hy+ provides a user interface with extensive support for visualizing structural (or relational) data as hygraphs (of which labelled graphs are a special case). Hygraphs, defined in [Con92], are a hybrid between Harel’s higraphs [Har88] and directed hypergraphs. Hygraphs constitute a convenient abstraction that generalizes several diagrammatic notations. The Hy+ system supports visualizations of the actual database instances, and not just diagrammatic representations of the database schema, Given the large volume of data that the system must present to the user, it is fundamental to provide her with two fundamental capabilities. The first one is the ability to define new relationships by using queries. This is pretty much the traditional way of using database queries: the newly defined relationship either gives a direct answer to a user question, or it provides a new view on the existing data. The derived data can later be presented visually by the system. The second capability is an innovative way of using queries to decide what data to show. Using this capability the user can selectively restrict the amount of information to be displayed, This filtering of relevant information is fundamental if one is to have any hope of conveying manageable volumes of visual information to the user. Selective data visualization can be used to locate relevant data and to restrict visualization to interesting portions of it (i.e., deciding what data to present), and can also be used to control the level of detail at which the data is presented (i.e., choosing how to see the data). To describe queries, Hy+ relies on a visual pattern-based notation. The patterns are expressions of the GraphLog query language [Con89, CM90b]. Overall, the system supports query visualization (i.e., presenting the description of the