Calculating cumulative percentage or percentage per group for each time can sometimes be a task with a slight twist. Let’s check this with ggplot2 and tidyverse.

library(ggplot2)
library(tidyverse)

data <- data.frame(
  sector = rep(1:20, each = 5),
  item = rep(1:5, times = 20),   
  value = rpois(100, 10)         
)

Three (out of many more) ways to show how this can be achieved.

using simple calculation

# using simple calculation
ggplot(data, aes(x = factor(sector), y = value / sum(value) * 100, fill = factor(item))) +
  geom_bar(stat = "identity", position = "fill") +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(x = "Sector", y = "Percentage", title = "Stacked 100% Bar Plot by Sector") + coord_flip()

replacing percent with tapply

# replacing percent with tapply
ggplot(data, aes(x = factor(sector), y = value / tapply(value, sector, sum)[as.character(sector)] * 100, fill = factor(item))) +
  geom_bar(stat = "identity", position = "fill") +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(x = "Sector", y = "Percentage", title = "Stacked 100% Bar Plot by Sector") + coord_flip()

without any complications

ggplot(data, aes(x = factor(sector), y = value, fill = factor(item))) +
  geom_bar(stat = "identity", position = "fill") +
  scale_y_continuous(labels = scales::percent_format(), name = "Percentage") +
  labs(x = "Sector", title = "Stacked 100% Bar Plot by Sector") + coord_flip()

And by all means, the diagram is in all cases the same. Just the examples can be slightly more over-engineered

Happy R-coding and stay healthy!