Coordinating declarative queries with a direct manipulation data exploration environment

Abstract

Interactive visualization techniques allow data exploration to be a continuous process, rather than a discrete sequence of queries and results as in traditional database systems. However limitations in expressive power of current visualization systems force users to go outside the system and form a new dataset in order to perform certain operations, such as those involving the relationship among multiple objects. Further, there is no support for integrating data from the new dataset into previous visualizations, so users must recreate them. Visage's information centric paradigm provides an architectural hook for linking data across multiple queries, removing this overhead. This paper describes the addition to Visage of a visual query language, called VQE, which allows users to express more complicated queries than in previous interactive visualization systems. Visualizations can be created from queries and vice versa. When either is updated, the other changes to maintain consistency.