Authoring Tools for Transparent Climate Reporting
summary
Built a prototype called Fluid transparent programming language that tracks data flow to make charts and supporting text queryable and provenance-visible. Reproduces a policy assessment with 61 risks drawn from an Excel sheet, exposing overlaps between views while only showing about ~5,000 elements with five-row slices. Text numbers become computed queries; runtime builds a dependency graph in about 10s dependency graph build. Known gaps: missing chapter labels in source data, coarse dependency analysis, performance and usability; roadmap: domain-native columnar ops, scoped/relative queries, LLM-based interpretation assistant.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:23Science vs communication gap
- 2:05Fluid transparent language
- 3:10Reproducing policy doc in Fluid
- 7:45Visualization interface complexity
- 8:05Code structure for visualizations
- 8:48Transparent text for reports
- 13:05Three ingredients for transparency
- 14:29Dependency analysis challenges
- 15:36Domain alignment for efficiency
- 16:41Exploring computation steps
- 17:01Intentional provenance
- 18:06Q&A
related talks
16mSometimes you just need words- 16mSometimes you just need words
19mSoftware Visualization: The humane solution
27mClear And Presentation Danger - How To Share Data-driven Insights
01h 30mCreating responsive data visualisations
01h 10mStorytelling through data
20mDesign for Local Government: Data, Services, and Transparency
45mContent First Prototyping