Toward Understanding Similarity of Visualization Techniques

TL;DR

This paper explores the similarity of visualization techniques using a model-driven approach based on technique signatures and an expert-driven assessment, providing initial insights into their relationships.

Contribution

It introduces a combined model-driven and expert-driven methodology to analyze visualization technique similarities, an area lacking comprehensive understanding.

Findings

Preliminary insights into visualization technique similarities

Comparison of model-based and expert-based assessments

Identification of relationships among 13 visualization techniques

Abstract

The literature describes many visualization techniques for different types of data, tasks, and application contexts, and new techniques are proposed on a regular basis. Visualization surveys try to capture the immense space of techniques and structure it with meaningful categorizations. Yet, it remains difficult to understand the similarity of visualization techniques in general. We approach this open research question from two angles. First, we follow a model-driven approach that is based on defining the signature of visualization techniques and interpreting the similarity of signatures as the similarity of their associated techniques. Second, following an expert-driven approach, we asked visualization experts in a small online study for their ad-hoc intuitive assessment of the similarity of pairs of visualization techniques. From both approaches, we gain insight into the similarity of a set of 13 basic and advanced visualizations for different types of data. While our results are so far preliminary and academic, they are first steps toward better understanding the similarity of visualization techniques.