Distinct 3D Glyphs with Data Layering for Highly Dense Multivariate Data Plots

TL;DR

This paper introduces a novel glyph design paradigm for dense multivariate data visualization, enabling clearer interpretation of complex datasets through layered visual encoding and distinct 3D shapes.

Contribution

It proposes a new glyph design approach that enhances discrimination and interpretability of dense multivariate plots using layered visual encoding and distinct 3D shapes.

Findings

Effective visualization of over 6 variables in a single plot

Improved discrimination of data categories in dense plots

Application success in high energy physics and security domains

Abstract

A carefully constructed scatterplot can reveal plenty about an underlying data set. However, in most cases visually mining and understanding a large multivariate data set requires more finesse, and greater level of interactivity to really grasp the full spectrum of the information being presented. We present a paradigm for glyph design and use in the creation of single plots presenting multiple variables of information. We center our design on two key concepts. The first concept is that visually it is easier to discriminate between completely distinct shapes rather than subtly different ones, specially when partially occluded. The second one is that users ingest information in layers, i.e. in an order of visual relevance. Using this paradigm, we present complex data as binned into desired and relevant discrete categories. We show results in the areas of high energy physics and security, displaying over 6 distinct data variables in each single plot, yielding a clear, highly readable, and effective visualization.