stacked barplot with aggregated data (ggplot2)

Question

I have some major problems with ggplot2. Even though it might be a very easy question to you I couldnt manage yet to get it right (I have read a ggplot2-book and was looking on stackoverflow as well).

Orginally there was a dataset consisting of a factor variable (country) and a dichotomous variable. Unfortunately I dont have my data in this extended format myself: I have two variables "var1" and "var2". var1 gives the number of cases in the original dataset where a certain condition is true and var2 gives the number of cases where the same condition is false:

country | var1 | var2
----------------------
"A" | 20 | 30
"B" | 24 | 53
"C" | 21 | 24
"D" | 30 | 66

Now I'd like to produce a stacked barplot with a y-axis showing percentages within each country (all bars should have the same height) plus the absolute numbers displayed within the bars:

How it should look

I found out that if the data were in an extended format, I could use

ggplot(data=dataset)+geom_bar(aes(x=country, fill=variable), position='fill')

However, I only have aggregated data.

Could anyone please help me?

Thank you!

Answer 1

A quick solution by first reshaping and then plotting:

library(ggplot2)
library(tidyr)

temp2 <- gather(temp, var, val, -country)
ggplot(temp2, aes(country, val, fill = var)) + 
  geom_col(position = 'fill') +
  geom_text(aes(label = val), position = position_fill(vjust = 0.5)) +
  scale_y_continuous(labels = scales::percent_format())

Answer 2

library('ggplot2')
library('data.table')

df1 <- fread('country var1 var2
             "A" 20 30
             "B" 24 53
             "C" 21 24
             "D" 30 66')

df1 <- melt(df1, id.vars = 'country', )
df1[, percent := prop.table(value)*100, by = 'country']

ggplot(data = df1, aes( x = country, y = percent, fill = variable, label = value )) + 
  geom_bar(stat = 'identity') + 
  geom_text(size = 5, position = position_stack(vjust = 0.5))