Data Visualization for Documents
Overview
The goal of a chart is insight, not decoration. Choose the right form for the relationship you want to show and design it so the message is immediate.
When to Use
- Adding charts to reports, decks, or dashboards
- Deciding between chart types
- Improving existing visualizations that are confusing or misleading
Chart Selection Guidance
- Comparison → bar / column
- Trend over time → line
- Part-to-whole → carefully used stacked bar or limited pie/donut
- Distribution → histogram or box plot
- Relationship → scatter
- Geographic → map (when location is essential)
Design Principles
- Start with the question the chart should answer
- Remove non-essential ink (heavy gridlines, unnecessary 3D, chartjunk)
- Label directly when possible instead of relying only on legends
- Use color with purpose (and check accessibility)
- Keep scales honest — avoid truncated axes that mislead unless clearly called out
- Prefer clarity over novelty
Verification
- The main insight is visible within a few seconds
- Chart type matches the data relationship
- Labels, units, and sources are present
- The visual does not distort the data