R Pie Chart with only top three segments

Question

I have a question regarding pie charts in R. I want to have a pie chart for used operating systems. However, there are many systems which have a share of under 1% and I want to exclude them from the diagram. Of course I could delete them from the dataframe before plotting, but I was wondering if there is a better alternative within the ggplot functions to plot only the top three operating systems.

In the following the dataframe as input and the code I am using:

Dataframe:

operatingSystem | sessionsPercent
Android         | 0.620
iOS             | 0.360
Windows         | 0.010
Blackberry      | 0.001
...

Code:

p <- ggplot(df, aes(x="", y=sessions, fill = operatingSystem)) 
p + geom_bar(width = 1, stat = "identity") +
  geom_text(aes(label = percent(data$sessions)), position = position_stack(vjust = 0.5), color = "white", size = 8) +
  coord_polar(theta="y", direction = -1) +
  theme_void() 

Does anybody has an idea?

Answer 1

There is no need to delete these rows from your dataframe, you could just run your ggplot commands over a subset of your data. BTW, I think you mean some OS could have a share under 1%, not under 0%? Also, please do not name a data.frame "data", because it could mess up some functions from the utils package. Better to use df for example.

I think you can try this:

library(ggplot2)
library(scales)

df <- read.table(text = 'operatingSystem  sessionsPercent
             Android          0.620
             iOS              0.360
             Windows          0.010
             Blackberry       0.001', header = TRUE)

p <- ggplot(df[df$sessionsPercent %in% head(df$sessionsPercent, 3),], aes(x="", y=sessionsPercent, fill = operatingSystem)) 

p + geom_bar(width = 1, stat = "identity") +
  geom_text(aes(label = percent(head(df$sessionsPercent, 3))), position = position_stack(vjust = 0.5), color = "white", size = 8) +
  coord_polar(theta="y", direction = -1) +
  theme_void() 

Percentages do not add up to 100% now, but if you want that, you could divide the argument inside the percent() command by 0.99 (since total percentage of 3 OS with highest percentages is 99%).