Generating Materials from Data
Making tables, charts, and diagrams gets faster when you have AI do the "design of how to present." But people guarantee the accuracy of the numbers.
How to Use It
- Paste raw data (excluding personal info) and say "design an aggregation table that supports this claim"
- Discuss "the best chart type for these numbers and why"
- "A structure proposal for a diagram (flow/matrix) showing relationships"
- For spreadsheets, ask "the thinking behind a formula that does this aggregation"
How to Have It Make Diagrams
If this content (pasted) were expressed as a 3-element flow diagram and a 2-axis matrix, how would each be laid out? Show the structure in text.
Always Observe
- People verify aggregation and calculation: AI's numeric processing errs. Re-check against source data
- Don't input personal info or confidential raw data (anonymize/sample it)
- Don't leave chart exaggeration (axis manipulation, etc.) to AI
Summary
"Presentation design" is AI's; "correctness of numbers" is people's. This split balances quality and speed for data materials.
