Pick the Right Chart for the Story You Want to Tell
Match a dataset and a question to the chart type that communicates clearly — including charts to avoid.
When to use this
When you have data and a question but every chart you try seems wrong.
The prompt
You are a data visualization designer who reads Edward Tufte and ignores chart libraries' defaults.
- **My data** (columns and a few rows or summary stats): [...]
- **The question I want the chart to answer**: [in one sentence]
- **Audience**: [who's going to look at it]
- **Where it'll appear**: [slide / dashboard / blog post / report]
Do this:
1. **The right chart, with reasoning** — recommend one chart type. Say WHY this matches the question and data. Don't just list options.
2. **Encoding choices** — what goes on each axis, color, size. Be explicit about what each visual variable encodes.
3. **One alternative chart** for the same data — a different chart that would answer a SLIGHTLY different question. Useful if I phrased my question wrong.
4. **Charts to AVOID** — name 1–2 chart types that look tempting for this data but would obscure the answer. Say why.
5. **The honest title** — propose a chart title that states the conclusion the chart supports. Not "Revenue Over Time" but "Revenue doubled in Q3".
6. **What to drop** — gridlines, legends, decimals, etc. Default chart libraries add too much.
What you'll get back
A primary chart recommendation with reasoning, an alternative chart for adjacent questions, charts to avoid, a conclusion-style title, and a list of defaults to strip.
How this is structured in English
Notice the English patterns this prompt uses — they're worth borrowing for your own requests.
- Ignores chart libraries' defaults Specific bias to push against. Default settings rarely produce the best chart; saying this out loud changes the AI's defaults.