Catalog

Guideline Catalog

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

781 records

Page 26 of 33

  1. Use a sequential colormap when values have no meaningful midpoint

    For ordered or continuous value comparison without a meaningful midpoint, use sequential color encoding on color-mapped charts to improve fidelity and mitigate false pattern detection for readers including those with color-vision deficiency.

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

  2. Use a shared line chart for same-date comparisons

    For same-date comparison over ordered time, use single-view structure on multi-series line charts to improve fidelity and mitigate cross-panel comparison errors for readers judging which series is higher at a specific point.

    • purpose:select
    • basis:heuristic
    • task:compare
    • time:ordered-time
    • structure:single-view:use
    • +3

  3. Use a shared-space layout for local time-point extrema

    For local extremum comparison in ordered time, prefer a shared-space layout on multi-series temporal charts to improve readability and mitigate slow cross-row scanning for readers doing quick overview analysis.

    • purpose:select
    • basis:empirical
    • task:extreme
    • time:ordered-time
    • structure:single-view:use
    • +3

  4. Use a single 24-hour bar chart to find the daily maximum

    For finding the maximum in cyclic-time hourly data, use a single-view structure on bar charts to improve fidelity and mitigate wrong-half selections for non-expert readers.

    • purpose:select
    • basis:empirical
    • task:extreme
    • time:cyclic-time
    • chart:bar
    • +4

  5. Use a single 24-hour linear bar chart for AM and PM interval comparisons

    For exact paired AM/PM comparison in cyclic-time hourly data, use a single-view structure on linear bar charts to improve fidelity and mitigate underestimation of small value differences for non-expert readers.

    • purpose:select
    • basis:empirical
    • task:compare
    • time:cyclic-time
    • chart:bar
    • +5

  6. Use a single panel instead of row facets for individual-value reading

    For exact value lookup and pairwise comparison, prefer a single-panel layout on point-based multivariate views to improve fidelity and mitigate slow cross-panel scanning for readers inspecting individual values.

    • purpose:select
    • basis:empirical
    • quality:fidelity:use
    • lever:layout-structure
    • operator:lookup
    • +3

  7. Use a single-view specification workspace for targeted follow-up questions

    For targeted follow-up analysis of tabular data, prefer a single-view specification workspace over a multi-view recommendation gallery to improve focused question answering and address cumbersome scanning for analysts.

    • purpose:select
    • basis:empirical
    • structure:single-view:use
    • structure:multi-view:avoid
    • data:tabular
    • +3

  8. Use a small discrete-outcome encoding to support uncertainty recall

    For short-term recall of uncertainty from a single reported experiment, prefer a small discrete-outcome encoding on uncertainty displays to improve fidelity and address weak memory for distribution shape among statistically novice readers.

    • purpose:refine
    • basis:empirical
    • scope:single-result
    • lever:encoding
    • operator:uncertainty
    • +3

  9. Use a stacked bar chart instead of a pie chart when there are more than five shares

    For part-whole reading on a single total with more than five shares, use a stacked bar chart instead of a pie chart to improve readability and mitigate untidy slice labeling for readers.

    • purpose:select
    • basis:heuristic
    • chart:bar:use
    • chart:pie-donut:avoid
    • lever:chart-family
    • +3

  10. Use a stacked bar chart instead of a stacked area chart for negative correlation judgments

    For relate tasks, use a stacked bar chart on negative-correlation displays instead of a stacked area chart to improve fidelity and mitigate imprecise discrimination of nearby association strengths for readers judging correlation.

    • purpose:select
    • basis:empirical
    • task:relate
    • chart:bar:use
    • chart:area:avoid
    • +3

  11. Use a stacked bar chart instead of a stacked line chart for negative correlation judgments

    For relate tasks, use a stacked bar chart on negative-correlation displays instead of a stacked line chart to improve fidelity and mitigate imprecise discrimination of nearby association strengths for readers judging correlation.

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

  12. Use a stacked bar chart to compare shares across two totals

    For comparing shares across two totals, use a stacked bar chart instead of a pie chart to improve readability and mitigate the single-total limit of pie charts for readers.

    • purpose:select
    • basis:heuristic
    • task:compare
    • scope:grouped-result
    • chart:bar:use
    • +4

  13. Use a stacked column chart when there are fewer than about ten dates

    For showing shares over ordered time with few dates, prefer a stacked column chart instead of an area chart to improve readability and mitigate hard labels and value reading for readers comparing several shares.

    • purpose:select
    • basis:heuristic
    • task:trend
    • time:ordered-time
    • chart:bar:use
    • +4

  14. Use a standard uncertainty display and explain it in text

    For communicating uncertain estimates, use explicit uncertainty encoding with textual explanation on charts that display statistical confidence or uncertainty to improve accessibility and mitigate ambiguous interpretation for readers who need unambiguous chart reading.

    • purpose:refine
    • basis:accessibility
    • quality:accessibility
    • lever:text-annotation
    • operator:uncertainty
    • +2

  15. Use a symbol map for absolute regional totals

    For comparison of non-temporal regional totals, prefer a symbol map over a choropleth on geospatial data to improve fidelity and mitigate misleading area-based comparison for readers interpreting raw counts.

    • purpose:select
    • basis:heuristic
    • task:compare
    • chart:map:use
    • chart:choropleth:avoid
    • +3

  16. Use a symbol map for many specific locations

    For showing many exact locations on a map, prefer a symbol map over a choropleth to improve fidelity and mitigate forcing point data into administrative regions for map readers.

    • purpose:select
    • basis:heuristic
    • chart:map:use
    • chart:choropleth:avoid
    • data:geospatial
    • +3

  17. Use a table for structured record lookup

    For retrieve tasks on structured record lists, use a table instead of text to improve lookup speed and mitigate missed reader-relevant details for readers seeking the entry that applies to them.

    • purpose:select
    • basis:heuristic
    • task:retrieve
    • chart:table:use
    • chart:text:avoid
    • +3

  18. Use a table instead of a line chart for exact value lookup

    For exact lookup tasks on ordered quantitative data, prefer a table over a line chart to improve fidelity and mitigate unnecessary visual inference for viewers retrieving specific values.

    • purpose:select
    • basis:empirical
    • task:retrieve
    • reading-mode:exact
    • chart:table:use
    • +3

  19. Use a table instead of a parallel coordinates plot for exact value lookup

    For exact value-retrieval tasks on small multivariate record sets, prefer a table over a parallel coordinates plot to improve fidelity and address slow cross-axis tracing for exact-reading situations.

    • purpose:select
    • basis:empirical
    • task:retrieve
    • chart:table:use
    • chart:parallel:avoid
    • +3

  20. Use a table instead of a scatter plot for exact value lookup

    For exact value-retrieval tasks on small multivariate record sets, prefer a table over a scatter plot to improve fidelity and address slow cross-attribute lookup for exact-reading situations.

    • purpose:select
    • basis:empirical
    • task:retrieve
    • chart:table:use
    • chart:scatter:avoid
    • +3

  21. Use a table when exact regional values or tiny regions matter

    For exact lookup and comparison of regional values, prefer a table over a choropleth on geospatial data to improve readability and mitigate color-based imprecision when subtle differences or very small regions matter.

    • purpose:select
    • basis:heuristic
    • chart:table:use
    • chart:choropleth:avoid
    • data:geospatial
    • +3

  22. Use a topology-preserving cartogram for adjacency lookup

    For lookup tasks on geospatial cartograms, prefer a topology-preserving cartogram type on distorted map views to improve fidelity and mitigate neighbor-identification errors for readers checking which regions touch.

    • purpose:refine
    • basis:empirical
    • task:relate
    • chart:map
    • data:geospatial
    • +3

  23. Use a treemap instead of a node-link tree when a weighted hierarchy must fit in one view

    For overview of large weighted hierarchies, use a space-filling treemap instead of a node-link tree on hierarchical data to improve insight and mitigate screen-space overload for users scanning many levels at once.

    • purpose:select
    • basis:empirical
    • chart:treemap:use
    • chart:network:avoid
    • data:hierarchical
    • +4

  24. Use a violin plot instead of a mean bar chart with error bars when distribution matters

    For grouped comparisons where distribution shape can change interpretation, use a violin plot on grouped results to improve insight and mitigate within-the-bar bias for analysts comparing sample populations.

    • purpose:select
    • basis:empirical
    • task:compare
    • scope:grouped-result
    • chart:box-violin:use
    • +3