SQL, data modeling, data quality, EDA, visualization · insight-1070
Replace rainbow palettes and many-color legends with ordered scales and direct labels
- Evidence type
- documented fact
- Difficulty
- easy
- Roles
- Data analyst, Data scientist
- Industries
- cross-industry
Summary: situation, action, reason, boundary
Situation
A route-by-month heat map uses a rainbow scale, and a line chart of eleven routes needs a legend to read.
Action
Use a sequential scale for ordered values and a diverging scale only around a meaningful midpoint; limit qualitative colors, label lines directly, add a second cue such as shape, and test with a color-vision simulator.
Reason
Lightness order carries magnitude without a legend, while redundant coding keeps the chart readable when color alone fails.
Boundary
Heat maps, maps, multi-series charts and dashboards shared with large audiences.
Symptom or risk
Rainbow lightness is non-monotonic, so arbitrary bands look like features and ends look similar; many categorical colors are hard to match, and some readers cannot separate red and green.
Exceptions
Categorical data with two or three groups, where simple color is enough.
Trade-offs
Direct labels need room and care to avoid overlap.
Pitfalls
Using red and green as the only difference between good and bad.
Lessons and concepts
Claims and their sources
Wilke describes the rainbow scale as non-monotonic in lightness, notes that roughly 8% of males and 0.5% of females have a color-vision deficiency, and advises limiting qualitative colors and testing with a simulator.
Wilke recommends redundant coding and direct labeling so that figures do not depend on color or a legend alone.
Diverging scales should have a neutral midpoint with real meaning, such as zero.