Euclidean, Hyperbolic, and Spherical Networks: An Empirical Study of Matching Network Structure to Best Visualizations
Loading...
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
The Eurographics Association and John Wiley & Sons Ltd.
Abstract
We investigate the usability of Euclidean, spherical and hyperbolic geometries for network visualization. Several techniques have been proposed for both spherical and hyperbolic network visualization tools, based on the fact that some networks admit lower embedding error (distortion) in such non-Euclidean geometries. However, it is not yet known whether a lower embedding error translates to human subject benefits, e.g., better task accuracy or lower task completion time. We design, implement, conduct, and analyze a human subjects study to compare Euclidean, spherical and hyperbolic network visualizations using tasks that span the network task taxonomy. While in some cases accuracy and response times are negatively impacted when using non-Euclidean visualizations, the evaluation shows that differences in accuracy for hyperbolic and spherical visualizations are not statistically significant when compared to Euclidean visualizations. Additionally, differences in response times for spherical visualizations are not statistically significant compared to Euclidean visualizations.
Description
CCS Concepts: Human-centered computing → Visualization design and evaluation methods; Empirical studies in visualization
@article{10.1111:cgf.70126,
journal = {Computer Graphics Forum},
title = {{Euclidean, Hyperbolic, and Spherical Networks: An Empirical Study of Matching Network Structure to Best Visualizations}},
author = {Miller, Jacob and Bhatia, Dhruv and Purchase, Helen and Kobourov, Stephen},
year = {2025},
publisher = {The Eurographics Association and John Wiley & Sons Ltd.},
ISSN = {1467-8659},
DOI = {10.1111/cgf.70126}
}