The Subspace Voyager: Exploring High-Dimensional Data along a Continuum of Salient 3D Subspaces

TL;DR

This paper introduces a visual exploration framework that decomposes high-dimensional data into a continuum of 3D subspaces, enabling analysts to interactively explore, navigate, and understand complex data structures.

Contribution

The framework allows interactive exploration of high-dimensional data through a continuum of 3D subspaces with data-driven tools for subspace selection and navigation, enhancing understanding beyond traditional methods.

Findings

Effective visualization of high-dimensional data via 3D subspace exploration.

Tools for guiding users to interesting subspaces and managing exploration.

Demonstrated applicability in clustering, information discovery, and classifier training.

Abstract

Analyzing high-dimensional data and finding hidden patterns is a difficult problem and has attracted numerous research efforts. Automated methods can be useful to some extent but bringing the data analyst into the loop via interactive visual tools can help the discovery process tremendously. An inherent problem in this effort is that humans lack the mental capacity to truly understand spaces exceeding three spatial dimensions. To keep within this limitation, we describe a framework that decomposes a high-dimensional data space into a continuum of generalized 3D subspaces. Analysts can then explore these 3D subspaces individually via the familiar trackball interface, but using additional facilities to smoothly transition to adjacent subspaces for expanded space comprehension. Since the number of such subspaces suffers from combinatorial explosion, we provide a set of data-driven subspace selection and navigation tools which can guide users to interesting subspaces and views. A subspace trail map allows users to manage the explored subspaces, and also helps them navigate within and across any higher-dimensional subspaces identified by clustering. Both trackball and trail map are each embedded into a word cloud of attribute labels, sized according to the relevance of the associated data dimensions in the currently selected subspace. Finally, a view gallery helps users keep their bearings and return to interesting subspaces and views. We demonstrate our system via several use cases in a diverse set of application areas, such as cluster analysis and refinement, information discovery, and supervised training of classifiers.