| name | swd-chart-guide |
| description | Storytelling with Data chart guidance. Use when choosing chart types, evaluating whether a chart communicates clearly, reviewing data visualizations for SWD compliance, proposing chart alternatives, or when the user asks "what chart should I use", "is this chart working", "how do I improve this chart", "show me a makeover example", or describes a visualization problem. |
| argument-hint | [chart type, anti-pattern, or description of visualization problem] |
Storytelling with Data (SWD) chart selection and decluttering guidance from Cole Nussbaumer Knaflic.
Core SWD Principles
The one insight rule: Every chart communicates exactly one thing. If you can't state the insight as a single declarative sentence, the chart is trying to do too much.
The 5-second test: A reader with no context identifies the main point within five seconds. If they can't, the visual hierarchy is wrong.
Color = meaning, not decoration: Start with gray for everything. Add color only to the element that carries the story. Never use color for categories unless each color encodes a specific meaning.
Declutter ruthlessly:
- Remove gridlines (or make them barely visible)
- Remove chart borders
- Remove unnecessary tick marks
- Move the legend into the chart as direct labels
- Never use 3D effects, shadows, or gradients
- Never use pie charts with more than 2-3 slices
Prose coherence: The title above the chart IS the insight sentence. If the title says "Revenue over time" but the point is "Revenue fell 40% after the policy change", the title is wrong. Titles should be declarative, not descriptive.
When to Use Which Chart
See chart-index.md for the full guide:
- When to use each chart type
- Common mistakes for each type
- Path to the full SWD guide for each chart
Quick reference:
| Use case | Chart type |
|---|
| Comparing categories at one point in time | Horizontal bar (sorted by value) |
| Showing change over time | Line chart |
| Showing composition / part of whole | Stacked bar (max 3-4 categories), or waffle |
| Comparing two groups across attributes | Dot plot or slopegraph |
| Showing before/after or two time points | Slopegraph |
| Distribution of a continuous variable | Histogram or dot strip |
| Relationship between two variables | Scatterplot |
| Sequential additions/subtractions | Waterfall chart |
| Flow between states | Sankey diagram |
Never use (and what to use instead):
- Pie chart with 4+ slices → horizontal bar sorted by value
- 3D bar or 3D pie → flat equivalent
- Dual y-axis → two separate charts or index to a common baseline
- Spider/radar chart → small multiples or dot plot
- Bubble chart for 3+ series → small multiples
Before/After Makeover Examples
See makeover-index.md for the full index.
When a chart has a specific problem, find the matching makeover example to show the before/after transformation. The index maps anti-patterns to real SWD makeovers.
Key examples:
- Pie chart → bar chart:
pie-chart-makeover, alternatives-to-pies
- Cluttered line chart → focused:
declutter-graphics, trade-forecast-makeover
- Bars vs. lines decision:
bars-vs-lines
- Dot plot vs. slopegraph vs. bar:
dot-plot-vs-slopegraph-vs-bar
- Stacked area → focused:
multitudes-of-makeovers, combo-chart-makeover
- Diverging bar problems:
diverging-bar-makeover
Full Reference Files
All content is hosted as public GitHub Gists. Use WebFetch to load any file on demand.
Chart guides gist (20 files):
https://gist.github.com/shakestzd/ecf5012aec9466110ca39371b9a527c7
Raw URL pattern: https://gist.githubusercontent.com/shakestzd/ecf5012aec9466110ca39371b9a527c7/raw/<filename>
Makeover examples gist (15 articles + YAML index):
https://gist.github.com/shakestzd/a4b156c29d82030236b40c1b6e02cf87
Raw URL pattern: https://gist.githubusercontent.com/shakestzd/a4b156c29d82030236b40c1b6e02cf87/raw/<filename>
Load any specific file on demand using WebFetch when deeper guidance is needed.
