Interactive Data Analysis: Visualization and Beyond
summary
Argues concrete design choices: uses Vega/Vega‑Lite, D3, Protovis and Alt and shows Voyager and Wrangler (now Trifacta) prototypes that generate runnable scripts (Python, Spark) from interactions. Reports Voyager users see 4× more variable sets and interact with 2× more unique plots. Presents a perceptual experiment where changing color maps and switching 3D to 2D raises diagnostic accuracy from ~40% to ~71%. Demonstrations include interactive previews, inferred transformations, and end-to-end export that actually works.
AI-written summary · about
chapters (17)
- 0:00Interactive data analysis intro
- 1:23Effective visualizations & interactions
- 4:32Building interactive tools
- 7:40What makes visualizations effective
- 10:26Graphical perception research
- 13:14Medical diagnostics redesign
- 15:41Broader data exploration
- 18:29Voyager 2 demo
- 23:46Voyager’s exploration impact
- 25:54Beyond visualization
- 27:21Data wrangling challenge
- 28:24Data Wrangler demo
- 34:19Commercializing wrangling tools
- 37:28Inverting analysis workflow
- 38:11Automation vs human control
- 40:38Future challenges in tools
- 41:00Closing remarks
related talks
10mThe Future of Data Visualization
32mData Viz for Engineers: Optimizing Insight & Decision Making Through Visualization
46mExplorations in Interactive Visual Analytics
20mVoyager 2: Augmenting Visual Analysis with Partial View Specifications
50mDVUX: Data Visualization and User Experience
31mFundamentals of Data Visualization
26m"Where Numbers Meet Art: Building Data Visualizations In The Modern Web"
20mearn to make the point: data visualisation strategy