Pie chart drawing in R-CodePudding

I want to draw a pie chart like below: Pie

A sample of the dataset is like this

structure(list(`Valuation Area` = c("20G5", "20G5", "20G5", "20G5", 
"20G5", "20G5", "20G5", "20G5", "20G5", "20G5", "20G5", "20G5", 
"20G5", "20G5", "20G5"), Backorder = c(15, 6, NA, 7, 5, 14, NA, 
NA, 4, NA, 3, NA, 12, NA, 1), `New higher` = c("yes", "yes", 
"yes", "yes", "yes", "yes", "yes", "no", "yes", "yes", "yes", 
"no", "no", "no", "yes")), row.names = c(NA, -15L), class = c("tbl_df", 
"tbl", "data.frame"))

My step 1,

library(dplyr)
library(tidyr)
library(ggplot2)

dput(Top_purchasing_2021[1:15,c(1,20,24)]) %>%
    drop_na(Backorder) %>%
    summarise(sum_backorder = sum(`Valuation Area` == "20G5"),
              sum_reduce_safetystock = sum(`New higher` == "yes"),
              sum_increase_safetystock = sum_backorder-sum_reduce_safetystock)

My step 2,

df <-  data.frame(group = c("sum_reduce_safetystock", "sum_increase_safetystock"), value = c(8, 1))
bp <-  ggplot(df, aes(x="", y=value, fill=group))  
  geom_bar(width=2, stat="identity")
bp
pie <- bp   coord_polar("y", start=0)   labs(title="backorder with safety stock")
pie

As can be seen "value=c(8,1)" is what i got from the step 1. when I change the code to below, it still draws the pie but shows error "Error in data.frame(group = c("sum_reduce_safetystock", "sum_increase_safetystock"), : object 'sum_reduce_safetystock' not found"

df <-  data.frame(group = c("sum_reduce_safetystock", "sum_increase_safetystock"), value = c(sum_reduce_safetystock, sum_increase_safetystock))
bp <-  ggplot(df, aes(x="", y=value, fill=group))  
  geom_bar(width=2, stat="identity")
bp
pie <- bp   coord_polar("y", start=0)   labs(title="backorder with safety stock")
pie

How can I fix this, or can i do it in a better way pls? thank you.

CodePudding user response：

ggplot2 can do all the calculations for you, as long as you have the data in long format - you don't need to summarise yourself.

I would simply create the two binary variables that you want to plot, and then let geom_bar do the calculations. ggarrange from ggpubr can then arrange the two charts together.

library(dplyr)
library(ggplot2)

# Create binary variables
Top_purchasing_2021 <- Top_purchasing_2021 |> 
  mutate(backorder = ifelse(is.na(Backorder),
                            "Without backorder",
                            "With backorder"),
         aim = ifelse(`New higher` == "yes",
                      "To reduce safety stock",
                      "To increase safety stock"))

# First pie chart
a <- Top_purchasing_2021 |> 
    ggplot(aes(x = "", y = backorder, fill = backorder))  
    geom_bar(stat = "identity", width = 1)  
    coord_polar("y", start = 0)  
    theme_void()  
    theme(legend.position = "bottom")

# Second pie chart
b <- Top_purchasing_2021 |>
    filter(backorder == "With backorder") |>
    ggplot(aes(x = "", y = aim, fill = aim))  
    geom_bar(stat = "identity", width = 1)  
    coord_polar("y", start = 0)  
    theme_void()  
    scale_fill_manual(values = c("coral4", "coral2"))  
    theme(legend.position = "bottom")

# Combine into one
ggpubr::ggarrange(a, b)

NB - pie charts are often not the best data visualisation tool (which is probably why they're not very straightforward to implement in ggplot). See here for more info.