Reverse stop bar

Question

I create a complex histogram using ggplot as follows:

plot_df <- df[!is.na(df$levels), ] ggplot(plot_df, aes(group)) + geom_bar(aes(fill = levels), position = "fill") 

Which gives me something like this:

How can I change the order of the stacked bars themselves, so that level 1 is at the bottom and level 5 is at the top of each column?

I saw several questions on this issue (for example, “ How to control the ordering of a stacked histogram using an identifier on ggplot2” ), and the general solution seems to be to reorder the data at this level, since the one that ggplot uses determines the order

So I tried reordering with dplyr:

plot_df <- df[!is.na(df$levels), ] %>% arrange(desc(levels))

However, the plot goes the same way. It also doesn't seem to matter if I sort in ascending or descending order.

Here is a reproducible example:

 group <- c(1,2,3,4, 1,2,3,4, 1,2,3,4, 1,2,3,4, 1,2,3,4, 1,2,3,4) levels <- c("1","1","1","1","2","2","2","2","3","3","3","3","4","4","4","4","5","5","5","5","1","1","1","1") plot_df <- data.frame(group, levels) ggplot(plot_df, aes(group)) + geom_bar(aes(fill = levels), position = "fill") 

Answer 1

The ggplot2 version 2.2.0 release ggplot2 on the Stacking panel suggest:

If you want to stack in reverse, try forcats::fct_rev()

 library(ggplot2) # version 2.2.1 used plot_df <- data.frame(group = rep(1:4, 6), levels = factor(c(rep(1:5, each = 4), rep(1, 4)))) ggplot(plot_df, aes(group, fill = forcats::fct_rev(levels))) + geom_bar(position = "fill") 

This is the original plot :

 ggplot(plot_df, aes(group, fill = levels)) + geom_bar(position = "fill") 

Or using position_fill(reverse = TRUE) , as suggested by alistaire in his comment :

 ggplot(plot_df, aes(group, fill = levels)) + geom_bar(position = position_fill(reverse = TRUE)) 

Note that the levels (colors) in the legend are not in the same order as in the stacked bars.