Use this to check a figure while you're working on it. The output identifies unmet criteria and gives you measurements to compare across drafts.
Arguments
- plot
A
ggplotobject.- width, height
Intended printed size in inches, used by the checks and measurements that depend on it. Defaults to 6.5 by 4.
- measure
Logical. Report the data-ink ratio and the data density, which are the slow part? Defaults to
TRUE. This governs only those two, which Tufte states no threshold for and the audit therefore doesn't grade. Every stated criterion is checked either way, so the count of violations means the same thing whichever you pass.
Value
An object of class tufte_audit: a tibble with one row per
check, whose status is "fail" for a stated criterion that's
not met, "pass" for one that's met, "report" for a
measurement Tufte gives no threshold for, and "skip" for a check
that couldn't run. The count of unmet criteria, a single integer, is
attached as the "violations" attribute.
Details
I've kept a distinction between principles with a stated criterion and those without one. Bars measured from zero and a lie factor between 0.95 and 1.05 can be checked against explicit rules. The audit also checks the advice that graphics tend toward the horizontal and that non-data ink be removed.
Tufte asks that the data-ink ratio be maximised "within reason" and that data density increase, but he doesn't give either a numerical target. Those measurements are reported without a pass or fail.
I haven't combined the results into a score. That would require me to decide how much each criterion matters. Read the individual results and check whether the suggested changes help someone understand your figure.
See also
tufte_principles(), which marks which principles carry
a stated criterion and which don't.
Examples
library(ggplot2)
tufte_audit(ggplot(mtcars, aes(wt, mpg)) + geom_point())
#>
#> ── Tufte audit ──
#>
#> At 6.5in x 4in: 3 stated criteria not met.
#>
#> ── Not met
#> ✖ The panel uses a #EBEBEBFF background fill. Try removing it and compare
#> whether the figure is easier to read.
#> Erase non-data ink - VDQI ch. 4
#> ✖ Minor gridlines are drawn. Consider whether readers need this level of detail
#> to estimate values.
#> Erase non-data ink - VDQI ch. 4
#> ✖ No caption is present. Add the data source with label_source() so readers can
#> check where the numbers came from.
#> Documentation - Beautiful Evidence ch. 6
#>
#> ── Measured, not graded
#> Tufte states a direction for these rather than a threshold. Read them against
#> another draft of the same figure.
#> • Data-ink ratio 0.06: an estimated 6% of the ink comes from data layers.
#> Compare drafts at the same dimensions; there's no target value.
#> • Data density 3.1 entries per square inch: 64 estimated entries over 20.8
#> square inches. Read this alongside the figure and entry count.
#> • 1 distinct colour in use. This is a count, not a verdict on whether the
#> colours help readers.
#> • One series in one panel. There's no series grouping to separate into small
#> multiples.
#>
#> ── Met
#> • No full panel border
#> • No pie chart
#> • Lie factor within Tufte's band
#> • No legend to decode
#> • No variable encoded twice
#> • Wider than it is tall
#> • Ink clears the WCAG contrast minimum
#> • Measured labels fit at the printed size