Understanding Bias in Perceiving Dimensionality Reduction Projections

TL;DR

This paper investigates how visual interestingness influences practitioners' preferences for dimensionality reduction projections, revealing a bias that affects structural faithfulness and proposing strategies to mitigate it.

Contribution

It identifies and explains the bias towards visually interesting projections over structurally faithful ones in dimensionality reduction.

Findings

Visual interestingness biases projection selection.

Bias increases with color labels and shorter viewing times.

Strategies can reduce perceptual bias in DR analysis.

Abstract

Selecting the dimensionality reduction technique that faithfully represents the structure is essential for reliable visual communication and analytics. In reality, however, practitioners favor projections for other attractions, such as aesthetics and visual saliency, over the projection's structural faithfulness, a bias we define as visual interestingness. In this research, we conduct a user study that (1) verifies the existence of such bias and (2) explains why the bias exists. Our study suggests that visual interestingness biases practitioners' preferences when selecting projections for analysis, and this bias intensifies with color-encoded labels and shorter exposure time. Based on our findings, we discuss strategies to mitigate bias in perceiving and interpreting DR projections.