Ranjan Maitra
Iowa State University

Three-dimensional Radial Visualization of High-dimensional Continuous and Discrete Datasets

We develop methodology for 3D radial visualization of high-dimensional datasets. Our display engine is called RadViz3D andextends the classic RadViz that visualizes  multivariate data in the2D plane by mapping every record to a point inside the unitcircle. The classic RadViz display has equally-spaced anchor points on theunit circle, with each of them associated with an attribute or feature of the dataset. RadViz3D obtains equi-spaced anchor points exactly for the five Platonic solids and approximately for the other cases via a  Fibonacci grid. We show that distributing anchor points at least approximately uniformly on the 3D unit sphere provides a better visualization

than in 2D. We also propose a Max-Ratio Projection (MRP) method that utilizes the group information in high dimensions to provide distinctive lower-dimensional linear projections that are then displayed using Radviz3D. Our methodology is extended todatasets with discrete and mixed features where a generalized distributional transform is used in conjunction with copula models before applying MRP and RadViz3D visualization.

This work is joint with (most done by) Fan Dai and Yifan Zhu of Iowa State University and is supported in part by the  National Institute of Biomedical Imaging and Bioengineering (NIBIB) of the National Institutes of Health (NIH)under its Award No. R21EB016212, and by the United States Department of Agriculture (USDA)/National Institute of Food and Agriculture (NIFA), Hatch project IOW03617. The content of this paper however is  solely the responsibility of the  authors and does not represent the official views of either the NIBIB, the NIH or the USDA.

Refreshments at 3:45pm in Snedecor 2101.