Sorting factors in multipanel plot in ggplot2 according to the first panel

Question

Is it possible to sort factors in a multipanel plot in ggplot2 according to the first panel? The first panel decides the order and the remaining panels follow that order.

Here is an example:

require(ggplot2)
set.seed(36)
xx<-data.frame(YEAR=rep(c("X","Y"), each=20),
               CLONE=rep(c("A","B","C","D","E"), each=4, 2),
               TREAT=rep(c("T1","T2","T3","C"), 10),
               VALUE=sample(c(1:10), 40, replace=T))

ggplot(xx, aes(x=CLONE, y=VALUE, fill=YEAR)) + 
  geom_bar(stat="identity", position="dodge") +
  facet_wrap(~TREAT)

Which gives me this plot:

Now I would like to sort CLONE based on the VALUE in YEAR X in a descending order (highest to lowest) but only for the Control (C panel). This order should then be maintained for T1, T2, and T3. By looking at the plot above, I want panel C sorted as CLONE C, B or D (both are 5), A and E. This order of CLONE should then be replicated for the remaining panels.

Answer 1

There's no easy way to do this right in ggplot since you have to reorder CLONE by 3 conditions, TREAT, YEAR and VALUE, otherwise forcats::fct_reorder2 could have been an option. Instead, extract the order of CLONE from the subset of data corresponding to YEAR = "X", TREAT = "C", and re-define your factor levels for the whole data set based on this subset.

library("ggplot2")
library("dplyr")
set.seed(36)

xx <- data.frame(YEAR = rep(c("X","Y"), each = 20),
           CLONE = rep(c("A","B","C","D","E"), each = 4, 2),
           TREAT = rep(c("T1","T2","T3","C"), 10),
           VALUE = sample(c(1:10), 40, replace = TRUE), stringsAsFactors = FALSE)

clone_order <- xx %>% subset(TREAT == "C"  & YEAR == "X") %>%
  arrange(-VALUE) %>% select(CLONE) %>% unlist()

xx <- xx %>% mutate(CLONE = factor(CLONE, levels = clone_order))

ggplot(xx, aes(x = CLONE, y = VALUE, fill = YEAR)) + 
  geom_bar(stat = "identity", position = "dodge") +
  facet_wrap(~TREAT)

giving