Specify label.y positions in facets about ggpubr HOT 3 CLOSED

Hi,

specifying label.y in aes() is not supported.

You can do this:

ggplot(test, aes(x=group, y = values, fill=group)) + 
  geom_boxplot() +
  geom_blank(aes(y=ymax*3))+
  facet_wrap(~vars, scales = 'free') + 
  stat_compare_means(comparisons=list(c("A","B")), label.y = 20)

When scales = 'free', it's even better to use label.y.npc:

ggplot(test, aes(x=group, y = values, fill=group)) + 
  geom_boxplot() +
  geom_blank(aes(y=ymax*3))+
  facet_wrap(~vars, scales = 'free') + 
  stat_compare_means(label.y.npc = 0.4)

Let's keep this issue open, I need to improve annotation of the different facet panels

from ggpubr.

Thanks @kassambara .
Surprisingly enough, label.y.npc does not work when I specify the comparisons.

This code

  geom_boxplot() +
  geom_blank(aes(y=ymax*3))+
  facet_wrap(~vars, scales = 'free') + 
  stat_compare_means(comparisons=list(c("A","B")),label.y.npc = 0.4)

actually still returns the labels at the top:

from ggpubr.

now, you can compute p-value and add it manually at any position.

In the following example, we'll use the ToothGrowth data. We'll compare supp levels after grouping the data by dose.

# Statistical test
library("rstatix")  # https://github.com/kassambara/rstatix
stat.test <-  ToothGrowth %>%
  group_by(dose) %>%
  t_test(len ~ supp) %>%
  adjust_pvalue() %>%
  add_significance("p.adj")
stat.test

# A tibble: 3 x 9
   dose   .y. group1 group2  statistic      p method  p.adj p.adj.signif
                           
1   0.5   len     OJ     VC  3.1697328 0.0064 T-test 0.0128            *
2   1.0   len     OJ     VC  4.0327696 0.0010 T-test 0.0030           **
3   2.0   len     OJ     VC -0.0461361 0.9600 T-test 0.9600           ns

# Visualization
library(ggpubr)
ggboxplot(
  ToothGrowth, x = "supp", y = "len",
  color = "supp", palette = "jco",
  ylim = c(0, 40)
) +
  facet_wrap(~dose) +
  stat_pvalue_manual(stat.test, label = "p.adj", y.position = 35)

Note that, the option y.position can be a numeric vector with length = the nombre of p-values to add. It can be also a column name in the stat.test data frame.

from ggpubr.