iSphere: Focus+Context Sphere Visualization for Interactive Large Graph Exploration
summary
Maps a 2D layout onto a Riemann sphere using inverse stereographic projection and renders that sphere back to the viewport; implements an object-based and pixel-based rendering pipeline (GPU pixel mapping for arbitrary visualizations). Evaluated against plane and hyperbolic displays on graphs with 128/512/2048 nodes (avg degree 8) across node, link, and path tasks. iSphere yields faster completion on link and path tasks—especially large graphs and small screens—hyperbolic is less accurate, users prefer iSphere for reduced interaction though curved links can hinder tracing.
AI-written summary · about
chapters (8)
- 0:00Introduction to the problem
- 1:00The sphere-based technique
- 2:07Projection and mapping
- 3:21Rendering approaches
- 4:28User study methodology
- 7:27Study results and feedback
- 10:04Window size evaluation
- 12:05Q&A and conclusion
related talks
01h 07mVisualizing Connections
46mExplorations in Interactive Visual Analytics
34mHigh-performance graphical view filtering with Sprotty
10mMovies Are in 3-D, Why Not our Data?...
42mVisualising Connections
24mCreating Interactive Data Visualizations for the Web
42mInteractive Data Analysis: Visualization and Beyond
13mState of the a-R-t - Advances in mapping in R