Releviz: Visualizing similar search results in two dimensions

We present Releviz, a system that provides a two-dimensional interface for selecting and displaying multivariate data records. Releviz describes a two-component system: A standardized constraint language for expressing search parameters specific to a particular data domain, and for evaluating those search parameters; and a two-dimensional display field of the search results, in linear or radial orientation, in which relevance maps to one of the major coordinate axes and a grouping function maps to the other. We provide two implemented grouping functions, weighted average, and a constraint-based clustering algorithm. We document and provide an internal evaluation of this system, including its sample implementation.