SPLOCIS - Extending SPLOMs to a Scatterplot Cube with Interactable Shadows for Immersive Analysis in Virtual Reality

Abstract

In data analysis, scatterplots serve as an initial tool for exploring the relationships between two or three attributes. While scatterplot matrices (SPLOMs) display every attribute combination through numerous 2D scatterplots to show a concise overview of a multivariate dataset, this approach is not directly suitable for 3D scatterplots due to visual clutter. Since research has shown that immersive virtual environments can enhance data analysis compared to traditional 2D desktop setups - especially for spatial analysis tasks - we propose an interactive system, called SPLOCIS, that makes use of virtual reality to enable users to interactively filter and select 3D scatterplots from all possible attribute combinations. Our user study, combining both qualitative and quantitative results, demonstrates that SPLOCIS is a particularly novel and stimulating approach to work with multivariate data in immersive environments. It enables solving classic data exploration tasks in an efficient and accurate way, while not imposing unexpectedly high task loads. Moreover, our findings provide promising suggestions for further developments.