每个标记的堆叠条的颜色不同？

Question

如何制作堆积条形图，使其每个堆积条的每个段都有不同的颜色（即，与所有条的总段一样多的唯一颜色 - 在本例中为 7 种不同颜色）。

我尝试了here 的方法，但由于输入数据的格式不同而得到了不同的结果，而且该问题显示了总数并且不需要图例（我需要图例）。

到目前为止的 MRE + 尝试

library(tidyverse)

df <- structure(list(discipline = c("Dev Ops", "Dev Ops", "Dev Ops", 
"Dev Ops", "Data Engineering", "Data Engineering", "Data Engineering"
), work_type = c("Casual/Vacation", "Contract/Temp", "Full Time", 
"Part Time", "Casual/Vacation", "Contract/Temp", "Full Time"), 
    n = c(3L, 117L, 581L, 9L, 1L, 297L, 490L)), class = c("tbl_df", 
"tbl", "data.frame"), row.names = c(NA, -7L))

# A tibble: 7 x 3
  discipline       work_type           n
  <chr>            <chr>           <int>
1 Dev Ops          Casual/Vacation     3
2 Dev Ops          Contract/Temp     117
3 Dev Ops          Full Time         581
4 Dev Ops          Part Time           9
5 Data Engineering Casual/Vacation     1
6 Data Engineering Contract/Temp     297
7 Data Engineering Full Time         490

这会生成正确的堆积条形图，但两个堆积条的颜色相同

df %>% 
  ggplot(aes(x = discipline, y = n, fill = work_type)) +
    geom_col(position = "Stack")

这会将唯一的颜色应用于每个堆叠条，但将相同应用于两个堆叠条

cols <- c("#5E4FA2", "#5E4FA2CC", "#5E4FA299", "#5E4FA266", "#9E0142", 
"#9E0142CC", "#9E014299")
df %>% 
  ggplot(aes(x = discipline, y = n, fill = work_type)) +
    geom_col(position = "Stack") +
  scale_fill_manual(values = cols[1:4])

这在两个堆叠条上实现了不同的颜色，但是颜色错误（和错误的图例）

df %>% 
  ggplot(aes(x = discipline, y = n, fill = cols)) +
    geom_col(position = "Stack")

这是基于this 方法，但请注意，条形高度与所有条形的总数匹配（而不是每个条形），并且在两个堆叠条形中也具有相同的颜色

df %>% 
  pivot_longer(cols = discipline:work_type) %>% 
  ggplot(aes(x = name, y = n)) + 
  geom_col(fill = c(cols, cols))

Answer 1

如果你想要两个结合两个因素，通常的技巧是使用interaction()，所以你的代码是：

# Data 
df <- structure(list(discipline = c("Dev Ops", "Dev Ops", "Dev Ops", 
"Dev Ops", "Data Engineering", "Data Engineering", "Data Engineering"
), work_type = c("Casual/Vacation", "Contract/Temp", "Full Time", 
"Part Time", "Casual/Vacation", "Contract/Temp", "Full Time"), 
    n = c(3L, 117L, 581L, 9L, 1L, 297L, 490L)), class = c("tbl_df", 
"tbl", "data.frame"), row.names = c(NA, -7L))

# Colours
cols <- c("#5E4FA2", "#5E4FA2CC", "#5E4FA299", "#5E4FA266", "#9E0142", 
"#9E0142CC", "#9E014299")

# Plot
df %>% 
  ggplot(aes(x = discipline, y = n, fill = interaction(work_type, discipline))) +
  geom_col(position = "Stack") +
  scale_fill_manual(name="Whatever", values = cols)

您可能想要一个更好的颜色图例名称，并且可能想要研究 interaction 的 sep 参数，以便这些因素更具可读性。