VISUALIZING LARGE, LOOSELY-STRUCTURED, HIERARCHICAL INFORMATION SPACES

Abstract

The magnitude of loosely, structured data available at users' fingertips from local and distributed sources is experiencing unprecedented growth. Even modern hierarchical visualization systems are overwhelmed by scalability problems due to the explosive increase in data magnitude. This thesis surveys the current state of the art in information visualization to determine essential visualization characteristics, leading to a set of guidelines for hierarchical visualization evaluation and design. The guidelines become the basis for a new hierarchicial visualization strategy, illustrated and implemented by the FLEX VIEW system. FLEX VIEW incorporates three main features of this strategy. First, the sophisticated use of emphasis and exclusion filters is applied to node attributes and an importance value to refine content quality. Second, the tight coupling of gestalt and detail views provides navigational capability and presents global trends and anomalies. Third, concepts are extended to abstracted hierarchies through automatic filter interface tuning. FLEX VIEW also introduces a \fIschlider\fR control for the compact selection of multiple, emphasized sub-ranges across a continuum.