Year: 2024
Type: conference-paper
Abstract: The scatterplot matrix is a widely used visualization model for exploring multi-dimensional data, as it allows for the simultaneous presentation of multiple scatterplots across various pairs of dimensions, providing an overview of potential data configurations. Nevertheless, understanding the relationships between multiple scatterplot units can be challenging. To address this issue, a scatter cube was introduced to facilitate interactive navigation among scatterplots. However, users need a high degree of spatial imagination to understand the transformation process between 2D scatterplots and 3D scatter cubes. To address this limitation, we present a novel approach called CubeFolding. In the transition animation of CubeFolding, we show how the scatter points in a scatter cube are obtained by orthogonal projection of scatter points in two related scatter plots. All the animations are done within the 2D scatterplot matrix overview without shifting the user’s gaze points. CubeFolding is adaptable for use in both desktop-based and immersive visualization environments. In addition, we conducted an empirical evaluation comparing the collapsed way with direct generation, and analyzed the effectiveness differences in helping users understand the scatter cube generation process between the two ways.
Authors: Shiran Liu, Meiyu Lv, Tiemeng Li
Institution: Beijing University of Posts and Telecommunications
Language: English
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4415685814