The Effect of Stereoscopic Immersive Environments on Projection-Based Multi-dimensional Data Visualization

Abstract

Multidimensional data impose a challenge for visual analyses. Commonly, dimensionality reduction techniques are used to project the multidimensional data into a 2D visual space. Poco et al. [9] showed that projection into a 3D visual space can increase the performance of common visual analysis tasks due to a higher projection precision. They also backed up their findings with a user study. However, when conducting the user study they displayed the 3D visual space on a 2D screen, which may impede the correct perception of the third dimension. In this paper, we present a study that investigates the effect of stereoscopic environments when used for the visual analysis of multidimensional data after projection into a 3D visual space. We conducted a controlled user study to compare correctness, timing, and confidence in segregation and precision tasks when performed in stereoscopic immersive environments and on a nonstereoscopic 2D screen. In terms of the stereoscopic immersive environments, we operated on and compared results obtained with two setup: a single screen and a six-sided highly immersive system, in both of which interaction was performed with a 3D input device. We investigated whether the stereoscopic immersive environments have an effect on user performance depending on the visual encodings. We used both 3D scatter plots and cluster visualizations in the form of enclosing surfaces or hulls for the visual analysis tasks.