Stacked geom_bar() with 2 bars per x value

Question

Forgive me is this is asked and answered, but so far I have not found a solution that meets my use case. If you know of one, please point me in the right direction.

Here is a small working sample of my code:

library(tidyverse)

source <- c("D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P")
subject <- c("M", "M", "M", "M", "M", "M", "M", "M", "M", "M", "M", "M", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R", "M", "M", "M", "M", "M", "M", "M", "M", "M", "M", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R")
grade <- c(1, 1, 1, 2, 2, 2, 1, 1, 1, 2, 2, 2, 1, 1, 1, 2, 2, 2, 1, 1, 1, 2, 2, 2, 1, 1, 2, 2, 2, 1, 1, 2, 2, 2, 1, 1, 2, 2, 2, 1, 1, 1, 2, 2, 2)
domain <- c("Alg", "Alg", "Alg", "Alg", "Alg", "Alg", "Geo", "Geo", "Geo", "Geo", "Geo", "Geo", "Comp", "Comp", "Comp", "Comp", "Comp", "Comp", "Voc", "Voc", "Voc", "Voc", "Voc", "Voc", "Alg", "Alg", "Alg", "Alg", "Alg", "Geo", "Geo", "Geo", "Geo", "Geo", "Comp", "Comp", "Comp", "Comp", "Comp", "Voc", "Voc", "Voc", "Voc", "Voc", "Voc")
placement <- c("A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A")
qty <- c(425, 389, 96, 460, 293, 163, 518, 291, 101, 366, 349, 201, 889, 661, 150, 680, 617, 465, 445, 293, 112, 381, 292, 208, 223, 232, 131, 270, 72, 27, 45, 9, 99, 40, 79, 194, 72, 126, 133, 123, 456, 98, 234, 432, 65)


test <- data.frame(source, subject, grade, domain, placement, qty)

plot4 <- test %>%
  ggplot(aes(x = grade, y = qty, fill = placement)) + 
  geom_bar(stat = "identity", position = "fill") +
  stat_count(aes(label = paste(sprintf("%1.2f", ..count../sum(..count..)*100), "%\n", ..count..), y = 0.5*..count..), 
                 geom = "text", 
                 colour = "black", 
                 size = 2.5, 
                 position = position_fill(vjust = 0.5)) +
  scale_x_discrete("Grade", limits = c(1, 2), 
                   labels = c("1st", "2nd"),
                   guide = guide_axis(angle = 90)) +
  scale_y_continuous("Prop place") +
  scale_fill_manual("Placement", 
                    values = rev(c("#de7e7e", "#ebeb4d", "#70e65e")),
                    labels = rev(c("C",
                                   "B",
                                   "A"))
  ) +
  theme(axis.title.y = element_blank(), 
        axis.text.y = element_blank()) +
  facet_wrap(vars(subject, domain), scales = "free_x")

plot4

which produces the following plot:

What I need for each grade level (in this case, 1st and 2nd) is to have two stacked bars with the data centered in each region of the stack. There should be one bar from each source ("D" and "P") associated with each grade level.

I do not know how to add second variable to the x-axis to create the bar from each source FOR EACH grade. I can either do grade, or source, but not both.

*BONUS POINTS for helping me figure out the stat_count line. I need the n() and % of each stack relative to the total for each source/grade/placement value. These numbers aren't even close, but this code has worked for me on less complicated plots.

"but so far I have not found a solution that meets my use case" - please kindly share links that have solutions that don't help and tell us why not. Don't let us re-invent the wheel. — tjebo, Jan 24 '21 at 11:46
I am voting to close this question for lack of clarity. Also, it's likely a duplicate to https://stackoverflow.com/q/12715635/7941188 — tjebo, Jan 24 '21 at 11:49

score 1 · Accepted Answer · edited Jan 25 '21 at 07:10

This may be what you are after

my_labels <- test %>%
group_by(grade, source,domain, subject) %>%
summarise(n = qty,p = qty/sum(qty)) %>%
mutate(lab = paste(n,"\n",sprintf("%1.2f",p*100),"%"),
x = interaction (grade,source)) # get the labels calculated and add at the end

The interaction() function helps to get combinations of grade and source for plotting

test %>%
ggplot(aes(x = interaction (grade,source), 
y = qty, fill = placement)) + 
geom_bar(stat = "identity", position ="fill") +
scale_x_discrete("Grade",
labels = c("1st D","2nd D","1st P","2nd P"), # add new X labels
guide = guide_axis(angle = 90)) +
scale_fill_manual("Placement", 
values = rev(c("#de7e7e", "#ebeb4d", "#70e65e")),
labels = rev(c("C","B","A"))) +
scale_y_continuous("Prop place") +
facet_wrap(vars(subject, domain), scales = "free_x") +
geom_text(data = my_labels, aes(label = lab,y = p),position = position_stack(vjust = 0.2)) # adjust position to get labels where you want.

Some comments here may help Adding percentage labels to ggplot when using stat_count

Thank you @e.matt. When I run your code above, I get the following error: `Error: Aesthetics must be either length 1 or the same as the data (180): x and fill`. When I comment out the geom_text(), it runs just fine, so I am thinking I am missing something in the geom_text() layer. `The interaction()` worked great for getting me side by side stacked bar charts. Thank you for that! — Adam Caudell, Jan 24 '21 at 22:38

Adam Caudell · Answer 2 · 2021-01-25T16:48:41.020

Thank you @e.matt. With your answer (and a little tweaking) I was able to get this exactly how I wanted it. I used mutate() to add a pcnt column, making the geom_text() layer a bit easier to work with. The interaction() function worked perfectly to give me two stacked and dodged bar charts per category (Grade in this case). Below is the modified solution code that I was able to make work out:

source <- c("D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P")
subject <- c("M", "M", "M", "M", "M", "M", "M", "M", "M", "M", "M", "M", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R", "M", "M", "M", "M", "M", "M", "M", "M", "M", "M", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R", "R")
grade <- c(1, 1, 1, 2, 2, 2, 1, 1, 1, 2, 2, 2, 1, 1, 1, 2, 2, 2, 1, 1, 1, 2, 2, 2, 1, 1, 2, 2, 2, 1, 1, 2, 2, 2, 1, 1, 2, 2, 2, 1, 1, 1, 2, 2, 2)
domain <- c("Alg", "Alg", "Alg", "Alg", "Alg", "Alg", "Geo", "Geo", "Geo", "Geo", "Geo", "Geo", "Comp", "Comp", "Comp", "Comp", "Comp", "Comp", "Voc", "Voc", "Voc", "Voc", "Voc", "Voc", "Alg", "Alg", "Alg", "Alg", "Alg", "Geo", "Geo", "Geo", "Geo", "Geo", "Comp", "Comp", "Comp", "Comp", "Comp", "Voc", "Voc", "Voc", "Voc", "Voc", "Voc")
placement <- c("A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A", "B", "A", "C", "B", "A", "C", "B", "A", "C", "B", "A")
qty <- c(425, 389, 96, 460, 293, 163, 518, 291, 101, 366, 349, 201, 889, 661, 150, 680, 617, 465, 445, 293, 112, 381, 292, 208, 223, 232, 131, 270, 72, 27, 45, 9, 99, 40, 79, 194, 72, 126, 133, 123, 456, 98, 234, 432, 65)
#pcnt was added to the dataframe using mutate() as shown below
pcnt <- pcnt <- c(46.7, 41.16, 11.31, 50.22, 30.08, 16.77, 56.92, 38.39, 12.38, 39.96, 32.92, 20.85, 56.04, 30.97, 16.74, 45.52, 32.07, 19.21, 52.75, 35.9, 16.38, 43.67, 19.5, 13.5, 52.29, 44.33, 10.37, 38.59, 41.58, 37.74, 47.29, 38.26, 10.65, 57.43, 22.92, 19.21, 53.41, 44.51, 9.94, 37.68, 33.44, 22.72, 48, 34.29, 7.8, 72.64, 29.01, 5.38, 52.35, 37.28, 14.27, 43.25, 40.39, 31.37, 38.05, 36.65, 27.7, 73.17, 50, 33.33, 29.41, 6.08, 60, 15.27, 10.7, 58.68, 7.22, 21.83, 16.94, 29.65, 51.99, 22.14, 50.34, 44.29, 22.9, 49.49, 21.75, 41.04, 39.82, 21.24, 78.06, 22.12, 12.5, 17.2, 15.56, 77.62, 10.13, 15.18, 37.73, 32.9, 17.36, 100)

test <- data.frame(source, subject, grade, domain, placement, qty)

test <- test %>%
  group_by(subject, grade, domain, placement) %>%
  mutate(pcnt = round(qty / sum(qty) * 100, 2)) %>%
  arrange(domain, desc(qty))

plot <- test %>% 
  ggplot(aes(x = interaction(source, grade), 
             y = qty, 
             fill = placement)) + 
  geom_bar(stat = "identity", position = "fill") +
  scale_x_discrete("Grade", 
                   labels = c("1.D", "1.P","2.D", "2.P"),
                   guide = guide_axis(angle = 90)) +
  scale_y_continuous("Prop place") +
  scale_fill_manual("Relative Placement", 
                    values = rev(c("#de7e7e", "#ebeb4d", "#70e65e")),
                    labels = rev(c("C", "B", "A"))
  ) +
  facet_wrap(vars(domain), scales = "free_x") +
  geom_text(size = 2.5, aes(label = paste(sprintf("%1.2f", pcnt), "%\n", qty), y = 0.5 * qty), 
            position = position_fill(vjust=0.5))

plot

Stacked geom_bar() with 2 bars per x value

2 Answers2