Guidelines
Suggest edit

Use a diverging color scale when the data have a meaningful midpoint

For comparison in quantitative color-encoded charts, use a diverging color scale on data with a meaningful midpoint to improve insight and address one-directional magnitude readings for readers interpreting values above and below that midpoint.

  • purpose:refine
  • basis:heuristic
  • task:compare
  • data:quantitative
  • quality:insight
  • lever:encoding
  • polish:palette
  • aesthetic:color:use

advice

Center the color scale on a meaningful midpoint

Use a diverging color scale when the quantitative data have a meaningful middle value. For example, center the scale on zero, 50%, the average or median, an agreed threshold such as a poverty line, or a target such as a revenue goal.

reason

Why a meaningful midpoint helps

A diverging scale makes the middle color carry meaning instead of acting as a decorative center. Readers can then see whether values fall above or below a reference value, not just where they sit on a single low-to-high ramp.

Mechanism: The split palette turns the midpoint into an explicit benchmark and makes both sides of that benchmark visible.

Evidence: The article recommends diverging colors when there is a meaningful middle value and names zero, 50%, the average or median, an agreed threshold, and a target as valid centers (Muth, 2021).

context

Use when the midpoint is part of the message

  • User Goal: Show whether values are above or below a reference value.
  • Task: Compare values by side of midpoint and by distance from it.
  • Data: Quantitative values with an explicit middle such as zero, 50%, median, threshold, or target.
  • Chart Setting: A chart or map already using a quantitative color scale.
  • Audience: Readers need to interpret the center as a real benchmark.
  • Success Criterion: Readers can state what the center color means and tell which values fall on each side.

exceptions

Do not use when the middle is arbitrary

Break it when: The data do not have a meaningful middle value and are better read as a one-directional low-to-high range. Why: A diverging center would impose a split the data do not justify.

costs

Costs of adding a midpoint split

Sacrifice: You give up some immediate light-to-dark simplicity.
Risk: Readers may wonder what the middle color means if the center is weak or unexplained.
Mitigation: Use a midpoint readers can name and recognize.

mistakes

Common midpoint failure

Mistake: Centering a diverging scale on an arbitrary middle that is not a real benchmark. Why it fails: Readers infer an above-versus-below meaning that the data do not support.

check

Check that the midpoint is real

Failure Sign: Reviewers cannot explain what the middle color stands for.
Quick Check: Ask, “What exact value is the midpoint?” If it is not a named benchmark such as zero, 50%, median, threshold, or target, do not use diverging.
Stronger Test: Compare a sequential version; if the diverging center adds no interpretable above/below meaning, keep the sequential scale.

fix

Fix an arbitrary diverging center

  • Set the center of the scale to an explicit benchmark.
  • Replace the diverging scale with a sequential light-to-dark scale if no benchmark exists.
  • Preview both sequential and diverging versions and keep the one whose midpoint has a clear meaning.

References

Muth, L. C. (2021). When to use sequential and when to use diverging color scales. https://www.datawrapper.de/blog/diverging-vs-sequential-color-scales