Skip to content
Design process
- Iterate over your designs, get feedback
Visualization design
- A chart must be as simple as possible, but not simpler
- Rainbow color maps are harmful
- Don’t use piecharts
- Correspondence principle: The same data should produce the same visualization
- Maxime data-ink ratio
- Overview first, zoom and filter, detail on demand (Shneiderman’s Information seeking mantra)
- Use interaction in visualization sparsely and cautionsly
- Don’t use more than six colors together
- Reduce clutter in Parallel coordinates by reordering axes
- Use adjacency matrices for dense networks
- Chose your colors for various types of color-blindness as well as printed in black+white
- Use several complementary visualizations to reduce clutter and complexity
- Avoid overuse of texture
- Highlight and label missing data
- Focus on the patterns in your data
Communication & presentation
- Message first, visualization design second.
- Don’t truncate axes
- Label your charts and axes
- Avoid overlapping labels and other occlusions
- Chose meaningful scales
- Add meaningful annotations carefully.
- Lie factor guideline: The representation of numbers, as physically measured on the surface of the graphic itself, should be directly proportional to the quantities represented.
- Don’t use the blow-appart effect
- Introduce and explain your visualization
- Know your audience
- Chose a meaningful chart type
- Avoid redundant information and redundant visual encodings
- Avoid 3D effects and encoding data in 3D volumes
- Careful embellishment can improve memorization if related to the patterns in the data
- Use line charts for continuous (e.g., temporal) values only
- Avoid circular charts
- Rely on simple and well-known visualization unless you have a reason
- Long lines are difficult to follow.