Catalog

Guideline Catalog

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

781 records

Page 16 of 33

  1. Place the title at the top of the visualization

    For recall-oriented reading of a single visualization, use top-placed title text on the chart to improve insight and address missed title content for readers scanning the figure directly.

    • purpose:refine
    • basis:empirical
    • task:retrieve
    • quality:insight
    • lever:text-annotation
    • +2

  2. Plot slope ratios on a log scale

    For grouped-result comparison of target-absent to target-present slope ratios, use log encoding on quantitative ratio plots to improve fidelity and mitigate positive-skew distortion for domain experts.

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

  3. Plot the difference as its own series when differences are the message

    For difference comparison over ordered time, use a direct difference series on line charts to improve judgment fidelity and mitigate gap-reading errors for readers tracking change between two series.

    • purpose:refine
    • basis:empirical
    • task:compare
    • time:ordered-time
    • chart:line
    • +3

  4. Prefer a bar chart over a line chart for average-position estimates

    For short-delay average-position judgments, prefer the bar chart family over the line chart family on single-series quantitative displays to improve fidelity and mitigate systematic bias for readers estimating averages from memory.

    • purpose:select
    • basis:empirical
    • chart:bar:use
    • chart:line:avoid
    • data:quantitative
    • +3

  5. Prefer a colorfield over a line chart for interval-average judgments

    For overview comparison of averages across fixed intervals in ordered time series, prefer color encoding on a heatmap-like series display over a line chart to improve judgment fidelity and mitigate peak-finding errors for viewers summarizing dense data.

    • purpose:select
    • basis:empirical
    • task:compare
    • time:ordered-time
    • chart:heatmap:use
    • +3

  6. Prefer a palette with visible color transitions for gradient comparison

    For side-by-side gradient comparison, prefer palette choices with visible color transitions on color-coded quantitative scalar fields to improve discrimination of net gradient differences and mitigate missed structural differences for viewers making overview judgments.

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

  7. Prefer a perceptually uniform multi-hue sequential palette over a rainbow palette

    For retrieve tasks, prefer a perceptually uniform multi-hue sequential color scale on quantitative color encodings to improve fidelity and mitigate slower, more error-prone relative-value judgments for viewers comparing values against a legend.

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

  8. Prefer a superimposed log-scaled shared view for multivariate time-series comparison

    For comparison tasks on ordered time, prefer a superimposed log-scaled shared view on multivariate line charts to improve comparison speed and mitigate slow cross-row reading for readers comparing several series.

    • purpose:select
    • basis:empirical
    • task:compare
    • time:ordered-time
    • structure:single-view:use
    • +4

  9. Prefer a table over a stacked bar chart for trial safety review

    For safety-review comparison tasks on grouped treatment-outcome data, prefer a table over a bar chart to improve detection fidelity and mitigate missed adverse-event associations for domain experts.

    • purpose:select
    • basis:empirical
    • task:compare
    • chart:table:use
    • chart:bar:avoid
    • +4

  10. Prefer a temporal transition when equally simple next-step options compete

    For choosing the next slide in a linear narrative sequence, prefer temporal transitions on multi-view visualization presentations to improve readability and mitigate less favored measure, dimension, or granularity jumps for audiences following a guided presentation.

    • purpose:refine
    • basis:empirical
    • task:compose
    • time:ordered-time
    • data:temporal
    • +3

  11. Prefer abstract uncertainty symbols for fast grouped comparison

    For grouped comparison of aggregate uncertainty across multiple point symbols, prefer abstract encoding on map-like grouped displays to improve readability and mitigate slower region-level judgments in simple non-interactive comparisons.

    • purpose:refine
    • basis:empirical
    • task:compare
    • scope:grouped-result
    • operator:uncertainty
    • +3

  12. Prefer bars over lines for delayed average-position judgments

    For average-position reports after a short delay, prefer bar charts over line charts on single-series quantitative displays to improve fidelity and mitigate systematic position bias for readers reproducing a series' average level from memory.

    • purpose:select
    • basis:empirical
    • chart:bar:use
    • chart:line:avoid
    • lever:chart-family
    • +2

  13. Prefer candidate choropleth maps with stronger regionalized spatial patterns

    For single-result thematic map selection, prefer variable encodings on choropleth maps with stronger regionalized spatial patterns to maximize visual interest and address flat or arbitrary-looking geographic distributions for readers scanning for explanatory context.

    • purpose:refine
    • basis:empirical
    • scope:single-result
    • chart:choropleth
    • data:geospatial
    • +3

  14. Prefer charts over tables for weakly polarized persuasive messages

    For persuasive communication about debated topics, prefer chart-based data presentation on statistical evidence displays to maximize attitude change and address no-change responses for readers with neutral or weakly polarized initial attitudes.

    • purpose:select
    • basis:empirical
    • chart:bar:use
    • chart:line:use
    • chart:table:avoid
    • +3

  15. Prefer color saturation over hue for order judgments

    For order judgments, prefer color-saturation over hue on quantitative marks in ordered one-dimensional sequences to improve fidelity and mitigate missed ordered patterns for viewers scanning static displays.

    • purpose:refine
    • basis:empirical
    • task:relate
    • quality:fidelity:use
    • lever:encoding
    • +3

  16. Prefer color-saturation over area for a secondary quantitative field during exact value reading

    For exact value lookup and pairwise comparison, prefer color-saturation for a secondary quantitative field on point-based multivariate views to improve fidelity and mitigate interference from varying mark area for readers inspecting individual values.

    • purpose:refine
    • basis:empirical
    • quality:fidelity:use
    • lever:encoding
    • operator:lookup
    • +3

  17. Prefer discriminability-weighted hue palettes for aggregate color comparisons

    For aggregate color comparisons, prefer a discriminability-weighted hue palette on nominal color encodings to improve fidelity and address confusion between similar categories for readers scanning many colored marks.

    • purpose:refine
    • basis:empirical
    • data:categorical
    • quality:fidelity:use
    • lever:encoding
    • +2

  18. Prefer filled point shapes when color differences must be read

    For point-to-point color comparison, prefer filled point shapes on scatterplots that encode information with both shape and color to improve fidelity and mitigate missed color differences for viewers reading multiclass points.

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

  19. Prefer mixed colors when topic associations should stay neutral

    For categorical palette design on charts and maps, prefer mixed rather than pure colors on data marks to prevent unwanted semantic cues and address colors that imply danger, goodness, or other loaded meanings for readers bringing cultural color associations.

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

  20. Prefer natural interpolation when clustered values and outliers need different treatment

    For regional comparison on outlier-rich value maps, prefer natural interpolation on choropleth color scales to improve fidelity and mitigate the tradeoff between washed-out linear scales and overstated equal-count scales for readers scanning geographic patterns.

    • purpose:refine
    • basis:heuristic
    • task:compare
    • chart:choropleth
    • data:geospatial
    • +3

  21. Prefer point lightness over point area for mean-position judgments in scatterplots

    For summary judgments of x/y mean position, prefer lightness encoding over point area on trivariate scatterplots to improve fidelity and mitigate weighted-average bias for readers deciding which points fall above or below average.

    • purpose:refine
    • basis:empirical
    • chart:scatter
    • measure:multi
    • lever:encoding
    • +3

  22. Prefer position encoding over color encoding for time-series extrema and range comparisons

    For point-level extremum and range comparisons in ordered time, prefer position encoding on time-series displays to maximize fidelity and mitigate missed interval highs and lows for exact interval judgments.

    • purpose:refine
    • basis:empirical
    • task:extreme
    • time:ordered-time
    • quality:fidelity
    • +3

  23. Prefer position over area for quantitative values

    For exact quantitative comparison, prefer position encoding on quantitative displays to improve reading accuracy and mitigate imprecise size judgments for viewers interpreting static graphics.

    • purpose:refine
    • basis:empirical
    • task:compare
    • data:quantitative
    • quality:fidelity:use
    • +4

  24. Prefer tables over charts for strongly opposed persuasive audiences

    For persuasive communication about debated topics, prefer tabular data presentation on statistical evidence displays to maximize attitude change and address no-change responses for readers with strongly negative initial attitudes.

    • purpose:select
    • basis:empirical
    • chart:table:use
    • chart:bar:avoid
    • chart:line:avoid
    • +3