具有动态返回值的 Ggplot scale_fill_manual

Question

I have a dataframe with multiple columns of factor type (with 5 levels: Dark Green, Green, Orange, Yellow, Grey) although each contains different values as per the screenshot below.

I tried to define a function to make a bar chart with colors of the legend manually set by the values contained in the selected column like so

library(tidyverse)
make_graph <- function(df, col){
  data = select(df, group, col)
  G = data$group
  C = data$col
  tbl = prop.table(table(G, C))
  DF = data.frame(tbl) %>% filter(freq != 0)
ggplot(data, aes(x=G,y=Freq, fill=C)) +
 geom_bar(stat="identity", position="dodge") + theme_minimal() +
 scale_fill_manual(values=c(?????), labels=c(?????))  # needs to be automatically filled with the values contained in the selected column

My question is: how can I set the parameters values and labels dynamically depending on the column provided to the function make_graph ? For example, if I do this make_graph(df, 'col_2') it will display a graph of 3 bars with colors dark green, green and orange and values of legend are 'great', 'good' and 'ok'

PS: below is the result of the code provided by akrun where the legend values were not aligned with the colors (for example darkgreen should be labelled as 'great' instead of 'terrible')

Answer 1

library(dplyr)
library(ggplot2)
make_graph <- function(df, col){
 keyval <- setNames(c("great", "good", "ok", "poor", "terrible"), 
      c("dark green", "green", "orange", "yellow", "grey"))
 fill_m <- keyval[df[[col]]]
df %>%
   select(group, all_of(col)) %>%
   count(across(everything())) %>%
   mutate(prop = n/sum(n)) %>%
   filter(n != 0) %>%
   ggplot(aes(x = group, y = n, fill = .data[[col]])) +
     geom_bar(stat= "identity", position = "dodge") + 
     theme_minimal() +
     scale_fill_manual(values = names(fill_m), labels = fill_m)



}

-testing

make_graph(df1, "col_2")

data

df1 <- structure(list(group = structure(c(1L, 2L, 1L, 2L), levels = c("1", 
"2"), class = "factor"), col_1 = c("dark green", "dark green", 
"dark green", "dark green"), col_2 = c("dark green", "green", 
"orange", "orange"), col_3 = c("green", "yellow", "dark green", 
"grey"), col_4 = c("grey", "dark green", "orange", "yellow"), 
    col_5 = c("dark green", "grey", "grey", "grey")), row.names = c(NA, 
-4L), class = c("tbl_df", "tbl", "data.frame"))