Guidelines
Suggest edit

Place the primary quantitative field on the x-axis in row-faceted point plots

For exact lookup and maximum search, prefer x-axis assignment for the primary quantitative field on row-faceted point plots to improve fidelity and mitigate slow misaligned cross-panel comparisons for readers scanning multiple panels.

  • purpose:refine
  • basis:empirical
  • quality:fidelity:use
  • lever:encoding
  • operator:lookup
  • reading-mode:exact
  • structure:small-multiples
  • channel:position:use

advice

Put the primary field on the x-axis across row facets

Place the primary quantitative field on the x-axis when categories are split into row facets and readers must read values or find maxima. For example, in a row-faceted point plot, swap the primary and secondary quantitative fields so the main field uses the common horizontal axis rather than the y-axis.

reason

Why x-axis assignment works better in row facets

The horizontal axis stays aligned across stacked panels, while the y-axis requires harder cross-panel lookups and can force the reader to scroll to the bottom to see the needed scale. That made the x-assigned version faster or more accurate in the studied row-faceted tasks.

Mechanism: A shared horizontal scale supports easier panel-to-panel comparison than repeated vertical lookups in stacked facets.

Evidence: Within the row-faceted designs, the version with the primary quantitative field on x was faster for read-value and compare-values tasks and more accurate for find-maximum than the transposed version with the primary field on y (Zeng & Battle, 2023; Kim & Heer, 2018).

Notes: This finding was reported for row facets specifically.

context

Use when the facet layout is already fixed

  • User Goal: Read exact values, compare individual values, or find the group containing the maximum value.
  • Task: Lookup, pairwise comparison, or maximum search.
  • Data: One categorical field split into row facets plus two quantitative fields.
  • Chart Setting: A row-faceted point plot is already chosen and the main quantitative field can be assigned to either x or y.
  • Success Criterion: Faster or more accurate reading across stacked panels.

exceptions

Do not generalize beyond the studied facet setup

Break it when: The layout is not row-faceted or the task is not one of the studied lookup or maximum tasks. Why: The reported advantage was specific to row facets and was not established as a general x-over-y rule for other layouts or tasks.

costs

What you give up

Sacrifice: The secondary quantitative field must take the y-axis. Risk: Leaving the primary field on y preserves the slower, harder cross-panel lookup pattern. Mitigation: Treat axis assignment as part of the facet design, not as a neutral swap.

mistakes

Common failure mode

Mistake: Keep the primary quantitative field on the y-axis in stacked row facets when readers must compare across panels. Why it fails: The repeated vertical lookups are less aligned and slower than reading from the shared horizontal axis.

check

Check the axis assignment directly

Failure Sign: Readers must compare y positions across stacked panels or scroll to find the primary scale. Quick Check: Transpose the two quantitative axes and compare one read-value question and one maximum-search question. Stronger Test: Measure time and error for the x-assigned and y-assigned versions on a small set of row-faceted tasks.

fix

Swap the axes inside the facets

  • Move the primary quantitative field from y to x.
  • Move the secondary quantitative field to y.
  • Keep the category field on row facets while re-testing the same tasks.

References

Kim, Y., & Heer, J. (2018). Assessing Effects of Task and Data Distribution on the Effectiveness of Visual Encodings. Computer Graphics Forum, 37(3), 157–167. https://doi.org/10.1111/cgf.13409
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