Supporting Data Journalism through Compilers for Visual Inputs
summary
Research demonstrates that visual graphics become code: reviz reverse engineers SVG subtrees from DOM into partial JavaScript programs with parameter holes (example: New York Times scatter retargeted to a penguin CSV) and compiles to observables plot in milliseconds instead of the hours or days manual reverse engineering takes. cartokit maps direct-manipulation edits into an IR and emits deployable Mapbox GL JS code (Washington Post choropleth reproduced and copy‑pasted to run).
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:08Data journalism vs compilers
- 5:24Data journalism challenges
- 6:48User study findings
- 8:33Compiling visual inputs
- 10:58Revis system demo
- 15:17How Revis works
- 19:55Revis benefits
- 21:39Cardit system demo
- 26:55Cardit interaction mapping
- 30:46Cardit workflow impact
- 31:51Compilers for visual inputs
- 34:42Graphics versus code
- 36:07Conclusion and future directions
related talks
29m"Hand Coded Artisanal Data-Visualization"
49mSaving the world with JavaScript: A Data Visualization story
38mThe Gamma: Programming Tools for Data Journalism
21mVisualizing data with JavaScript at Twitter
01h 55mUse your data to tell stories
47mData, Design, Code
01h 30mCreating responsive data visualisations
42mInteractive Data Analysis: Visualization and Beyond