Linked Visualisations via Galois Dependencies
summary
Presents a concrete, language-based dynamic analysis that links parts of visual outputs back to the input by computing Galois-style forward/backward slices, extended with De Morgan dual negation so selections can express “what competes for the same data.” Demonstrates brushing-and-linking on bar/line charts and a matrix-convolution demo, implements the idea in a prototype runtime and a small language called fluid, proves union-preservation and round-tripping, runs slowly as a prototype but artifacts are available and evaluated.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 0:11What are we interested in
- 1:01Brushing and Linking
- 4:31Use Case 2
- 7:56Set ourselves
- 11:43Represent selections
- 13:00Analysis
- 14:57DeMorgan Jewels
- 15:37Compositionality
- 16:26Motivation
- 16:58Applications
- 19:43Audience Question
related talks
19mFluid: towards transparent, self-explanatory research outputs
42mVisualizing Connections
01h 07mVisualizing Connections
30mVisualization Driven Development
34mVisualizations in Golang
34mVisualising Multi Dimensional Data
37mVisualizing Persistent Data Structures
18mGraphScape: A Model for Automated Reasoning about Visualization Similarity and Sequencing