PivotPaths: Strolling through Faceted Information Spaces

Area of Science:

• Information Visualization
• Human-Computer Interaction
• Data Exploration

Background:

• Modern information resources feature multiple facets and relations (e.g., academic publications, movies).
• Current navigation methods like filtering cause abrupt data subset changes, hindering comprehension of facet relationships and filter effects.

Purpose of the Study:

• To introduce PivotPaths, an interactive visualization system for exploring faceted information resources.
• To enable intuitive navigation and understanding of complex, interlinked data through visual paths and gradual transitions.

Main Methods:

• Developed PivotPaths with pivot operations for lightweight, gradual view transitions.
• Employed an iterative design-and-evaluation process.
• Conducted semi-structured interviews and a two-week deployment with a large academic publication database.

Main Results:

• PivotPaths exposes faceted relations as visual paths, facilitating a 'stroll' through information spaces.
• The interface supports gradual transitions, aiding comprehension of filter operations and facet relationships.
• User evaluation indicated an aesthetically pleasing design encouraging exploration and serendipitous discoveries.

Conclusions:

• PivotPaths provides an effective and engaging method for exploring large, faceted information collections.
• The visualization supports casual traversal and enhances the discovery of relationships within complex datasets.