Use discrete bins for bivariate color lookup
For exact lookup tasks on bivariate color maps, prefer discrete encoding on value-and-uncertainty charts to improve identification accuracy and mitigate perceptual color-decoding errors for readers matching marks to a legend.
- purpose:refine
- basis:empirical
- task:retrieve
- quality:fidelity:use
- lever:encoding
- operator:uncertainty
- reading-mode:exact
advice
Discretize the bivariate color scale
Quantize the value-and-uncertainty scale into explicit color bins when readers must identify exact pairs from a legend. For example, use a stepped 2D color matrix instead of a smooth bivariate gradient when a reader must locate a cell with a specific value and uncertainty.
reason
Why discrete bins beat continuous bivariate color for lookup
A continuous bivariate color field asks readers to estimate two integrated color dimensions precisely at once. Discrete bins trade some quantization error for easier matching between the map and the legend.
Mechanism: Discrete outputs reduce perceptual decoding error because readers compare a small set of distinct colors instead of estimating exact positions in a smooth 2D color blend.
Evidence: In the paper’s identification experiment, discrete bivariate maps were significantly more accurate than continuous bivariate maps for finding target value-uncertainty combinations (Correll et al., 2018).
context
Use when exact legend matching matters
- User Goal: Identify or verify a specific value-uncertainty combination.
- Task: Exact lookup from a bivariate legend.
- Data: Two quantitative variables encoded together in color, with one variable representing uncertainty.
- Chart Setting: A superimposed bivariate color map with a legend that readers must decode directly.
- Audience: Readers performing precise read-off rather than broad visual impression.
- Success Criterion: Higher accuracy in matching map colors to legend entries.
exceptions
Do not use when continuous appearance is the higher priority
Break it when: The main goal is preserving class-less continuous appearance rather than exact pair lookup. Why: The paper notes that discretization introduces quantization error even though it improves lookup performance.
costs
What you give up with discretization
Sacrifice: You lose continuous, class-less representation. Risk: Bins can hide small within-bin differences. Mitigation: Use discrete bins specifically for tasks that require exact matching from the legend.
mistakes
Common failure in bivariate color maps
Mistake: Use a smooth bivariate gradient and expect readers to recover precise value-uncertainty pairs from color alone. Why it fails: Precise estimation from integrated color channels is error-prone.
check
Test whether readers can actually match the colors
Failure Sign: Readers hesitate between neighboring colors or repeatedly miss exact target pairs. Quick Check: Ask a reviewer to find a specific value-uncertainty pair in the continuous version and in a discretized version of the same map. Stronger Test: Compare accuracy on a short identification task that uses the same legend and data in both versions.
fix
Revise the scale for exact reading
- Replace the smooth bivariate gradient with a stepped set of color bins.
- Turn the legend into explicit discrete categories rather than continuous interpolation.
- If exact lookup is not the chart’s job, keep the continuous scale out of tasks that ask for precise value-uncertainty identification.