Guidelines
Suggest edit

Use a color-encoded field instead of a line chart for aggregate time-range comparison

For aggregate comparison in ordered time, prefer a color-encoded field over a line chart on single-series temporal displays to improve accuracy and address detail-by-detail reading for viewers making overview judgments.

  • purpose:select
  • basis:empirical
  • task:compare
  • time:ordered-time
  • chart:heatmap:use
  • chart:line:avoid
  • quality:fidelity:use
  • lever:chart-family

advice

Choose a colorfield for aggregate range comparison

Use a color-encoded field instead of a line chart when readers need to compare averages across time ranges in a single series. For example, replace a single time-series line graph with a colorfield that maps value to color across the same ordered time axis when readers must pick the interval with the highest average.

reason

Why the colorfield works for this task

A color-encoded field supports interval-level reading as an overview task, while a line chart asks readers to inspect many individual positions and mentally combine them into range averages.

Mechanism: Mapping value to color across ordered time helps readers judge average level over a range more directly than summing many point heights from a line.

Evidence: In the collated experiment, the colorfield design ranked above the line chart for the aggregate task, and the accuracy difference was significant (ANOVA, p < .001) (Zeng & Battle, 2023; Correll et al., 2012).

Notes: The reported result is about accuracy on an aggregate judgment task, not about precise point lookup.

context

Use when the chart must support range-level averaging

  • User Goal: Identify which time interval has the highest average value.
  • Task: Compare aggregate values across fixed ranges within one ordered series.
  • Data: One quantitative series laid out over ordered time with many values per interval.
  • Chart Setting: Static single-view display of the same series.
  • Audience: Viewers making an overview judgment from the chart.
  • Success Criterion: Higher accuracy in choosing the correct interval.

exceptions

Do not use when the chart must support exact point reading

Break it when: The main task is reading exact individual values or inspecting detailed local patterns in the series. Why: The source positions the line graph as the standard display for detail-oriented reading, while the colorfield result is only supported for aggregate judgments.

costs

Tradeoffs of replacing the line chart

Sacrifice: You give up the position-based encoding used by the line chart. Risk: The color-encoded field can be the wrong choice if readers need exact values rather than interval averages. Mitigation: Keep the line chart when the task changes from range-level averaging to detailed value reading.

mistakes

Common failure mode in this decision

Mistake: Keeping a line chart for an average-over-ranges decision just because time series are conventionally shown as lines. Why it fails: The studied aggregate task was performed more accurately with the color-encoded field than with the line chart.

check

Check the choice with a direct comparison task

Failure Sign: Reviewers must visually trace many ups and downs in the line to decide which interval has the highest average. Quick Check: Show the same series once as the current line chart and once as a color-encoded field, then ask which interval has the highest average. Stronger Test: Run a small forced-choice test on matched series data and compare accuracy between the line chart and the color-encoded field.

fix

Fix the chart selection

  • Replace the line chart with a color-encoded field for the aggregate comparison task.
  • Keep the same ordered time axis so the compared intervals stay aligned.
  • Encode the quantitative value with color across the full series instead of position alone.
  • Switch back to a line chart if the task changes to precise point reading.

References

Correll, M., Albers, D., Franconeri, S., & Gleicher, M. (2012). Comparing averages in time series data. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 1095–1104. https://doi.org/10.1145/2207676.2208556
Zeng, Z., & Battle, L. (2023). A Review and Collation of Graphical Perception Knowledge for Visualization Recommendation. Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, 1–16. https://doi.org/10.1145/3544548.3581349