Replace rainbow colormaps with a perceptually ordered quantitative ramp
For comparison tasks on quantitative color scales, avoid rainbow colormaps on scalar color encodings to prevent slow and error-prone value reading and address misleading similarity judgments for viewers comparing relative differences.
- purpose:refine
- basis:empirical
- task:compare
- data:quantitative
- quality:fidelity
- lever:encoding
- operator:difference
- aesthetic:color:avoid
advice
Rainbow ramp replacement
Replace a rainbow colormap with a perceptually ordered quantitative ramp when color encodes ordered magnitude. For example, swap jet for viridis on a continuous heatmap legend; the tested sequential and perceptually uniform multi-hue ramps both outperformed jet overall.
reason
Why the rainbow ramp fails
A rainbow ramp changes hue in ways that do not consistently support ordered distance, so readers spend longer judging similarity and make more mistakes across the scale.
Mechanism: When hue boundaries and perceptual spacing do not align with numeric spacing, nearby values can look farther apart than larger value differences, which slows comparison and increases error.
Evidence: Jet was the slowest and most error-prone colormap tested overall, and the paper concludes that it should be jettisoned for quantitative color encoding despite one narrow high-performing region created by lucky color-name boundaries (Liu & Heer, 2018).
Notes: The one strong pocket for jet occurred where color-name boundaries happened to line up with the value grouping.
context
When this applies
- User Goal: Compare ordered values or gradients by color.
- Task: Judge which value is closer, more similar, or farther apart on a quantitative scale.
- Data: Ordered quantitative values encoded with a continuous color ramp.
- Chart Setting: A chart with a color legend where readers interpret the ramp as ordered magnitude.
- Audience: Viewers reading relative differences directly from color.
- Success Criterion: Faster and more accurate comparison across the full scale.
exceptions
When not to use it
Break it when: The encoding is intentionally discrete and relies on category-like color-name boundaries instead of a continuous ordered scale. Why: The only strong-performing jet cases were driven by categorical color-name transitions, not by reliable continuous quantitative reading.
costs
Costs of replacing the rainbow ramp
Sacrifice: You lose the vivid spectral look that often motivates rainbow use. Risk: Replacing jet with an arbitrary non-rainbow ramp can still leave comparison problems. Mitigation: Prefer a perceptually ordered ramp, especially a perceptually uniform multi-hue ramp, rather than any multi-hue palette.
mistakes
Common mistake with rainbow replacement
Mistake: Replacing jet with another multi-hue ramp that still lacks clear perceptual ordering. Why it fails: Multiple hues alone did not guarantee good performance; the tested multi-hue ramps worked when they were judiciously designed.
check
How to check it
Failure Sign: Comparisons slow down or become inconsistent around hue transitions. Quick Check: Sample low, mid, and high reference colors from the ramp and test a few triplets around major hue changes. Stronger Test: Compare the same chart in jet and in a perceptually ordered ramp such as viridis and keep the version that yields faster and more accurate similarity judgments.
fix
What to change
- Replace jet with a perceptually ordered ramp such as viridis.
- Re-test comparisons near major hue transitions after the swap.
- If the chart only needs a few discrete bins, simplify to a small sequential scale instead of using a rainbow ramp.