Catalog

Guideline Catalog

Browse visualization guideline records with sections, labels, and references.

781 records

Page 5 of 33

  1. Avoid assigning strong blue to Europe or white categories

    For non-temporal comparison of racial or regional groups, avoid strong blue category colors on white or Europe labels in categorical palettes to prevent superiority cues and mitigate competence, professionalism, or royalty associations for readers of race- or region-focused graphics.

    • purpose:refine
    • basis:heuristic
    • data:categorical
    • quality:trust
    • lever:encoding
    • +3

  2. Avoid brown and olive hues for racial categories

    For non-temporal comparison of racial or ethnic groups, avoid brown and olive category colors on categorical palettes to prevent unwanted skin-tone reminders and mitigate dislike of the encoded groups for readers of people-focused graphics.

    • purpose:refine
    • basis:heuristic
    • data:categorical
    • quality:trust
    • lever:encoding
    • +3

  3. Avoid color saturation as the uncertainty scale on point symbols

    For ordinal uncertainty reading on discrete point-symbol displays, avoid color saturation as the uncertainty encoding to prevent illogical symbol ordering and address misread certain-to-uncertain scales in simple non-interactive displays.

    • purpose:refine
    • basis:empirical
    • data:ordinal
    • operator:uncertainty
    • lever:encoding
    • +2

  4. Avoid colormaps that appear to vary in opacity when the background may change

    For quantitative comparison across changing presentation backgrounds, avoid opacity-varying color scales on colormap visualizations to improve reading robustness and mitigate background-dependent reversals for viewers interpreting magnitude from color.

    • purpose:refine
    • basis:empirical
    • quality:readability:use
    • lever:encoding
    • channel:opacity:avoid
    • +1

  5. Avoid dark yellowish-green colors when optimizing categorical palette preference

    For aesthetic evaluation of categorical palettes, avoid dark yellowish-green hues on categorical color encodings to maximize aesthetics and mitigate disliked color-region choices for viewers judging palette appeal.

    • purpose:refine
    • basis:empirical
    • data:categorical
    • quality:aesthetics
    • lever:encoding
    • +3

  6. Avoid dense visual structure for lay audiences

    For public-facing interpretation tasks, avoid dense visual structure on charts and multi-part displays to improve readability and mitigate confusion from technical-looking or cluttered presentations for lay viewers.

    • purpose:refine
    • basis:rhetorical
    • lever:layout-structure
    • density:dense:avoid
    • quality:readability:use
    • +3

  7. Avoid emotionally charged color choices when color encodes meaning

    For interpretation and comparison tasks, avoid emotionally charged color choices on charts where color encodes categories, scenarios, timepoints, or regional values to improve trust and mitigate mistaken color associations for lay audiences who rely on color as an early cue.

    • purpose:refine
    • basis:rhetorical
    • quality:trust
    • lever:encoding
    • communication:credibility
    • +3

  8. Avoid exact-point readouts from individual ensemble paths

    For exact location judgments on geospatial uncertainty maps, avoid using individual ensemble-member paths as the readout to prevent overweighting single members and mitigate biased point decisions for novice audiences.

    • purpose:refine
    • basis:empirical
    • chart:map
    • data:geospatial
    • quality:fidelity
    • +4

  9. Avoid extreme data-ink maximization in bar charts

    For subjective evaluation of simple quantitative comparisons, avoid extreme data-ink maximization on bar charts to improve aesthetics and mitigate low acceptance of unfamiliar minimalist forms for general-public audiences.

    • purpose:refine
    • basis:empirical
    • chart:bar
    • quality:aesthetics
    • lever:encoding
    • +2

  10. Avoid filled area encoding when viewers must estimate bivariate trends

    For by-eye trend estimation in bivariate charts, avoid filled area encoding on quantitative plots with ordered x-values to improve fidelity and mitigate less accurate relationship judgments for viewers inferring trends directly from the plotted data.

    • purpose:refine
    • basis:empirical
    • task:trend
    • data:quantitative
    • quality:fidelity:use
    • +2

  11. Avoid full-range context when values near 50% must be read without bias

    For exact single-value reading of part-to-whole data, avoid full-range contextual encoding on bar or dot charts to prevent midpoint bias and mitigate systematic underestimation below 50% and overestimation above 50% for readers judging values near the middle of the scale.

    • purpose:refine
    • basis:empirical
    • task:retrieve
    • quality:fidelity:use
    • lever:encoding
    • +2

  12. Avoid gray for 'other' or multiracial categories

    For non-temporal comparison of racial or ethnic groups, avoid gray category colors on 'other' or multiracial series in categorical charts to prevent diminished visual importance and mitigate signals that some groups matter less for readers represented in the data.

    • purpose:refine
    • basis:heuristic
    • data:categorical
    • quality:trust
    • lever:encoding
    • +2

  13. Avoid hard boundaries when encoding continuous uncertainty

    For uncertainty judgments on continuous data, avoid hard boundary encodings on maps or interval displays to improve fidelity and mitigate containment and deterministic reading errors for viewers interpreting probabilistic information.

    • purpose:refine
    • basis:empirical
    • operator:uncertainty
    • quality:fidelity:use
    • lever:encoding
    • +1

  14. Avoid hue-only encoding for ordered values

    For ordinal or quantitative reading, avoid hue-only encoding on charts with ordered values to improve fidelity and mitigate unordered magnitude readings for readers.

    • purpose:refine
    • basis:empirical
    • data:ordinal
    • data:quantitative
    • quality:fidelity:use
    • +2

  15. Avoid lightness differences when viewers compare numerosity

    For overview quantity comparisons, avoid lightness encoding on grouped marks to prevent biased numerosity judgments and mitigate false differences when viewers compare how many items belong to each group.

    • purpose:refine
    • basis:empirical
    • task:compare
    • quality:fidelity:use
    • lever:encoding
    • +2

  16. Avoid locking chart access to one technology path

    For digital chart access and interaction, avoid interaction-access that depends on a single browser, device, software environment, operating system, or input mechanism on charts to prevent fragile technology support and mitigate isolation of chart information or functionality for users of compliant assistive technologies.

    • purpose:refine
    • basis:accessibility
    • quality:accessibility
    • lever:interaction-access

  17. Avoid luminance encoding for group comparisons of numerosity

    For relative numerosity comparison, avoid luminance encoding for group identity on point-based displays to prevent count bias and mitigate darker-looks-more errors for overview readers.

    • purpose:refine
    • basis:empirical
    • task:compare
    • quality:fidelity
    • lever:encoding
    • +2

  18. Avoid opacity-varying colormaps when the same visualization must work on multiple backgrounds

    For aggregate comparison across changing display backgrounds, avoid opacity-varying palette construction on quantitative colormap views to improve encoding fidelity and mitigate background-dependent meaning shifts for viewers reading the same legend on light and dark presentations.

    • purpose:refine
    • basis:empirical
    • task:compare
    • data:quantitative
    • quality:fidelity:use
    • +3

  19. Avoid perfect-square targets when readers must compare rectangle areas

    For exact area comparison in area-based displays, avoid 1:1 aspect ratios on compared rectangles to improve fidelity and mitigate side-length heuristics for web viewers.

    • purpose:refine
    • basis:empirical
    • task:compare
    • chart:treemap
    • quality:fidelity
    • +3

  20. Avoid plotting two series in the same display for average-position estimates

    For short-delay average-position judgments, avoid multi-series layout structure on quantitative line and bar displays to prevent perceptual pull and mitigate biased average-position estimates for readers estimating averages from memory.

    • purpose:refine
    • basis:empirical
    • data:quantitative
    • quality:fidelity
    • lever:layout-structure
    • +3

  21. Avoid random scatter when comparing multiple icon arrays

    For first-glance comparison of multiple proportions under brief viewing, avoid random arrangement on icon-array graphics to improve fidelity and address rank confusion between moderately different values for public audiences.

    • purpose:refine
    • basis:empirical
    • task:compare
    • structure:multi-view
    • operator:difference
    • +3

  22. Avoid red and green pairs with similar brightness

    For categorical palette design intended for broad readership, avoid red and green category colors with similar brightness on data marks to improve accessibility and mitigate category confusion for readers with color vision deficiency.

    • purpose:refine
    • basis:heuristic
    • data:categorical
    • quality:accessibility
    • lever:encoding
    • +3

  23. Avoid seizure-triggering flashes and red-dominant visuals

    For static or active chart presentation, avoid flashing and red-dominant visual changes on chart content to prevent seizure risk and mitigate unsafe red flashes, red animation, and large red display areas for viewers with photosensitive seizure sensitivity.

    • purpose:refine
    • basis:accessibility
    • quality:accessibility
    • lever:interaction-access
    • access:motion-safe:use
    • +1

  24. Avoid slope encodings for pairwise relation judgments

    For compare tasks on grouped paired quantitative displays, avoid slope encoding on paired-value charts to improve fidelity and mitigate slow and inaccurate relation judgments for viewers.

    • purpose:refine
    • basis:empirical
    • task:compare
    • scope:grouped-result
    • data:quantitative
    • +3