distinguish groups and facets

[CorradoLanera]

Describe the solution you'd like

Suppose you have data like the following one

requireNamespace("tidyr")
requireNamespace("dplyr")
library(ggstatsplot)

sample_data <- tidyr::crossing(
  time     = factor(1:3),
  group    = letters[7:8],
  type     = LETTERS[1:4],
  value_id = 1:100
) %>% 
  dplyr::mutate(value = runif(row_number()))

I really love to obtain an enriched {ggstatsplot} for something like this:

sample_data %>% 
  ggplot(aes(x = time, y = value, colour = group)) +
  geom_violin() +
  facet_wrap(type ~ .)

My first (not working) attempt was

grouped_ggbetweenstats(
  data = sample_data,
  x = time,
  y = value,
  group = group, # {{!! I miss this one !!}}
  grouping.var = type,
  title.prefix = "Type",
  type = "np",
  pairwise.comparisons = TRUE, # between groups/value_id each time
  pairwise.annotation = "p.value",
  message = FALSE,
  title.text = "Differences in values by month (against groups) each type."
)

How could we do it?

I do not know if there is a workaround to achieve the expected result.
If I have to throw an idea, the only one I have in mind is not reliable (because it brakes to many things inside the package conventions):

Include the option grouping.var in ggbetweenstats() for groups (with comparisons made between groups) rename grouped_ggbetweenstats(), to facet_ggbetweenstats() including the option facet.var.

Note

I am not able to achieve it neither if I do not have "types": suppose there is a single type (ie no facet), how can I show the variation of groups comparison along time? If this is possible (and I have missed it), next, I have no problem to produce a distinct plot each type.

sample_data %>% 
  ggplot(aes(x = time, y = value, colour = group)) +
  geom_violin()

[CorradoLanera]

This could show the expected result (and why it is suboptimal :-))

sample_data <- tidyr::crossing(
  time     = 1:3,
  group    = factor(letters[7:8]),
  type     = factor(LETTERS[1:4]),
  value_id = 1:100
) %>%
  dplyr::mutate(value = runif(dplyr::row_number()))


p_list <- sample_data %>%
dplyr::group_by(.data = ., type) %>%
  dplyr::group_map(
    .f = ~ grouped_ggbetweenstats(
      data = .,
      x = group,
      y = value,
      plotgrid.args = list(
        align = "h",
        ncol = length(unique(sample_data[["time"]]))
      ),
      grouping.var = time,
      title.prefix = "time",
      type = "np",
      pairwise.comparisons = TRUE, # between groups/value_id each time
      pairwise.annotation = "p.value",
      message = FALSE
    )
  )

# combining plots
cowplot::plot_grid(
  plotlist = p_list,
  labels = levels(sample_data[["type"]]),
  align = "v",
  nrow = nlevels(sample_data[["type"]])
)

[IndrajeetPatil]

Closing in favor of #500