ETF5922 Chap.6 Multivariable Encodings and Small Multiples
Multivariable Encodings and Small Multiples
Channels have unequal precision
Position carries the two primary quantitative variables in a scatter plot. Qualitative, sequential and diverging palettes match different variable semantics.
Facets support repeated comparison
Facet wrap repeats a plot by one category, while facet grid organises panels by row and column variables.
Fixed facet scales support magnitude comparison; free scales expose local shape at the cost of direct comparison.
Accessibility as an analytical test
Keep facet scales fixed when the reader must compare magnitudes across panels.
Free scales can reveal within-panel shape, but they remove the common visual ruler and therefore change the question being answered.
Accessibility as an analytical test in context
Multivariable work is constrained by unequal perceptual accuracy. Position on a common scale supports fine quantitative comparison; size and area are coarser; hue is well suited to category identity but weak for ordered magnitude.
Small multiples trade a crowded combined view for repeated panels, provided axes remain comparable when magnitude across groups is the task. Accessible design is part of analytical correctness because an unavailable distinction is lost information.
The complete design argument names the variable carried by each channel, the judgment readers must make and the fallback that preserves that judgment under grayscale, small output or colour-vision differences.
What this chapter covers
- 01
Position carries the two primary quantitative variables in a scatter plot.
- 02
Qualitative, sequential and diverging palettes match different variable semantics.
- 03
Facet wrap repeats a plot by one category, while facet grid organises panels by row and column variables.
- 04
Fixed facet scales support magnitude comparison; free scales expose local shape at the cost of direct comparison.
- 05
Use position and colour to frame the reader's task
- 06
Check size against label before styling
- 07
Explain how palette changes the visible comparison
- 08
Audit sequential without removing necessary context
Assign channels without overloading the view
- 1Reserve position for the group means and calculate the nine-unit range.
- 3Encode status with an accessible categorical channel rather than a misleading area comparison.
- 2If panels are used, retain a common scale for cross-group magnitude.
- 2Remove any extra channel that does not represent a variable.
Key terms
- Visual channel
- A perceptual property such as position, hue, size or shape used to carry a variable.
- Small multiple
- A repeated panel using the same graphical design for different subsets of the data.
- Fixed scale
- A shared axis range that permits magnitude comparisons across panels.
- Accessible palette
- A set of colours distinguishable under common colour-vision differences and realistic reproduction.
Multivariable Encodings and Small Multiples FAQ
When should facet scales remain fixed?
Keep facet scales fixed when the reader must compare magnitudes across panels. Free scales can reveal within-panel shape, but they remove the common visual ruler and therefore change the question being answered.
How many variables should one visual encode?
There is no universal number. Add a channel only when it answers a named reader question and remains distinguishable at the output size. Position usually supports precise magnitude, while hue or shape can identify categories. Facets can reduce overload when panels retain comparable scales and labels.
Why test a palette in grayscale?
Grayscale exposes whether category identity or emphasis depends entirely on hue. If distinctions vanish, shape, direct labels, lightness or panel structure may provide a backup. The aim is not identical appearance; it is preservation of the comparisons the audience must make.
What should a small-multiple caption state?
State the subset represented by each panel, whether axes are fixed or free, and the comparison the arrangement supports. If scales differ, warn readers not to compare panel magnitudes directly. A caption can also name the channel used inside panels so repeated colour or shape meanings remain consistent across the complete set.
Exam move
Use one dataset with a quantitative outcome, a category and a status flag. Assign position, hue and shape, then remove each channel in turn to see which comparison disappears. Repeat with facets under fixed and free scales and inspect the page in grayscale. The exercise should identify both channel overload and the conditions under which small multiples remain comparable.
Working through Multivariable Encodings and Small Multiples in ETF5922? Sia is AskSia’s AI Computer Science tutor — ask any ETF5922 Multivariable Encodings and Small Multiples question and get a clear, step-by-step explanation grounded in how ETF5922 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
