stat_compare_means ¶

Test the groups of y along a discrete x and label the p-values.

As stat_compare_means in the R package ggpubr. Without comparisons or ref_group, one global test across the x categories is labelled at the top left ("Kruskal-Wallis, p = 1.5e-09"); when other aesthetics (color) split each x into groups, those groups are tested per x. With ref_group, every other category is tested against the reference and labelled above it. With comparisons, each pair gets a bracket (geom_bracket) stacked above the data, labelled with its p-value.

Parameters:

• mapping (FeatureSpec, default: None ) –

  Aesthetic mapping from aes(): x (categories) and y.

• data (DataFrame or dict, default: None ) –

  Layer data; inherits the plot data when None.

• method ((test, test, anova, test), default: "t.test" ) –

  The test: Welch's t test, Wilcoxon rank sum (signed rank when paired), one-way ANOVA or Kruskal-Wallis. Default "wilcox.test" for two groups (and for comparisons and ref_group), "kruskal.test" for more.

• paired (bool, default: False ) –

  Paired t or Wilcoxon tests; values pair up in row order.

• comparisons (list of pairs, default: None ) –

  Pairs of x categories to compare, by name or 1-based position, e.g. [("0.5", "1"), ("1", "2")].

• ref_group (str or int, default: None ) –

  Reference category, or ".all." to compare each category with all values.

• hide_ns (bool, default: False ) –

  With label="p.signif", hide ns (not significant) marks.

• label ((format, signif), default: "p.format" ) –

  "p.format": the p-value; "p.signif": stars (**** <= 1e-4, *** <= 0.001, ** <= 0.01, * <= 0.05, else ns). Default: the method and p-value ("Wilcoxon, p = 0.064"); for brackets, the p-value.

• label_sep (str, default: ", " ) –

  Separator between the method and the p-value.

• label_x (float, default: None ) –

  x of the global label, in data units.

• label_y (float or list of float, default: None ) –

  y of the labels, or the base of the brackets (one per comparison).

• step_increase (float, default: None ) –

  Gap between stacked brackets, as a share of the y range: 0.12, or 0 when label_y is given.

• tip_length (float, default: 0.03 ) –

  Bracket tip length, as a share of the y range.

• p_adjust_method (Literal, default: None ) –

  Adjust the layer's p-values for multiple comparisons (R p.adjust): "holm", "hochberg", "hommel", "bonferroni", "BH" (or "fdr"), "BY" or "none". The default "none" shows unadjusted p-values, as ggpubr.

• geom (str, default: None ) –

  "text" for labels, "bracket" for comparisons (the defaults).

• position (str or FeatureSpec, default: None ) –

  Position adjustment.

• show_legend (bool, default: None ) –

  True shows this layer in the legend.

• inherit_aes (bool, default: True ) –

  False ignores the plot's default mapping.

• manual_key (str or FeatureSpec, default: None ) –

  This layer's key in a manual legend.

• na_rm (bool, default: None ) –

  True removes rows with missing values silently.

• **other_args (Any) –

  Constant aesthetics of the labels and brackets, e.g. size, color.

Returns:

• LayerSpec –

  The layer.

Raises:

• SpecError –

  If a name is not accepted or a comparison is not a pair.

• UnsupportedValueError –

  If an option has the wrong type.

Examples:

>>> from ggrust import *
>>> data = {"diet": ["veg"] * 6 + ["meat"] * 6 + ["fish"] * 6,
...         "weight": [61, 58, 66, 70, 63, 59, 72, 80, 77, 69, 85, 74,
...                    65, 68, 71, 64, 70, 66]}
>>> plot = (ggplot(data, aes("diet", "weight")) + geom_boxplot()
...         + stat_compare_means(comparisons=[("veg", "meat"), ("veg", "fish")]))