Customizing colors in geom_arc_bar (ggforce)

Question

I am trying to plot scatter plot of pie charts using ggplot2 and ggforce. Following is my code

library(data.table)
library(ggplot2)
library(ggforce)
library(dplyr)

df <- fread("data_frame_pie_plot.csv", header = TRUE, 
        data.table = FALSE, stringsAsFactors = FALSE)

req_data <- apply(df, 1, function(x){
  geom_arc_bar(aes(x0 = x0, y0 = y0, r0 = 0, r = 0.45, 
               amount = amount, fill = discrete_fill), stat = "pie")})

ggplot(df) + req_data + theme_no_axes()

plot_obtained Link_to_data Data_frame.
Can I fill the colours manually (discrete(based on discrete_fill column) and continous (based on conti_fill column))? I have tried the following codes.

##================================================================================##
# Code 1
my_colors <- c("red","white","blue")
ggplot(df, fill = sample_mut_colour) + req_data +
  theme_no_axes() + scale_color_manual(values=my_colors)

##================================================================================##
# Code 2
my_colors_1 <- c("red", "white")
my_colors_2 <- c("blue", "white")

req_data <- apply(df, 1, function(x){
  geom_arc_bar(aes(x0 = x0, y0 = y0, r0 = 0, r = 0.45, 
               amount = amount, 
               fill = discrete_fill,
               color = case_when(discrete_fill == 1 ~ my_colors_1,
                                 discrete_fill == 2 ~ my_colors_2)), stat = "pie")})

ggplot(df) + req_data + theme_no_axes() 

##================================================================================##                  
# Code 3
req_data <- apply(df, 1, function(x){
  geom_arc_bar(aes(x0 = x0, y0 = y0, r0 = 0, r = 0.45, 
               amount = amount, 
               fill = case_when(discrete_fill == 1 ~ my_colors_1,
                                 discrete_fill == 2 ~ my_colors_2)), stat = "pie")})

ggplot(df) + req_data + theme_no_axes() 

##================================================================================##             
# Code 4
req_data <- apply(df, 1, function(x){
  geom_arc_bar(aes(x0 = x0, y0 = y0, r0 = 0, r = 0.45, 
               amount = amount, 
               fill = discrete_fill),
           palette = case_when(discrete_fill == 1 ~ "red",
                               discrete_fill == 2 ~ "black",
                               TRUE ~ "white"), stat = "pie")})

ggplot(df) + req_data + theme_no_axes()  

##=================================================================================##

But they do not seem to work. Any idea what is wrong in the code?

Answer 1

For a discrete scale your Code 1 example was nearly there. Just use scale_fill_manual instead of scale_color_manual. For more control over how the discrete values map to color values you can name the vector, too.

my_colors <- c("red","white","blue")
ggplot(df) + req_data +
    theme_no_axes() + scale_fill_manual(values=my_colors)
# or, if you want consistent output independent of the order of
# df$discrete_fill:
my_colors <- c("1" = "red", "2" = "white", "dummy" = "blue")
ggplot(df) + req_data +
    theme_no_axes() + scale_fill_manual(values=my_colors)

You can also just use a single call to geom_arc_bar for the whole data frame (rather than a list of call results from each row) and you'll get the same effect.

ggplot(df) +
  geom_arc_bar(aes(x0 = x0, y0 = y0, r0 = 0, r = 0.45, 
               amount = amount, fill = discrete_fill), stat = "pie") +
    theme_no_axes() + scale_fill_manual(values=my_colors)