在ggplot2中为箱线图与叠加散点设置不同边框颜色的方法

要实现箱线图边框和散点边框分别对应指定分组颜色，这里提供两种实用方案：

方案一：通过数据列绑定颜色（无需额外包） #

先给数据框添加两个颜色映射列，分别对应箱线图和散点的颜色：

library(tidyverse)
set.seed(1234)
df = tribble(
    ~group,~value,
    "a",   rnorm(1000, mean = 5),
    "b",   rnorm(1000, mean = 7, sd = 1.5),
    "c",   rnorm(1000, mean = 8),
    "c",   rnorm(1000, mean = 9),
    "c",   rnorm(1000, mean = 7)
) %>%
    unnest(value) %>%
    mutate(
        # 箱线图对应颜色
        box_color = case_match(
            group,
            "a" ~ "black",
            "b" ~ "blue",
            "c" ~ "brown"
        ),
        # 散点边框对应颜色
        point_color = case_match(
            group,
            "a" ~ "red",
            "b" ~ "green",
            "c" ~ "yellow"
        )
    )

绘图时直接调用这两个颜色列，并用scale_color_identity()告诉ggplot直接使用列中的颜色值：

library(ggplot2)
ggplot(df, aes(x = group, y = value)) +
    geom_boxplot(aes(color = box_color), fill = NA, outlier.shape = NA) +
    geom_point(aes(color = point_color), shape = 21, 
               position = position_jitter(seed = 1, width = .05), 
               alpha = 0.1) +
    scale_color_identity()

方案二：使用ggnewscale创建独立颜色刻度（无需修改原始数据） #

如果不想改动原始数据，可以用ggnewscale包给不同图层设置独立的颜色映射规则：

先安装并加载包：

install.packages("ggnewscale")
library(ggnewscale)

然后编写绘图代码：

library(ggplot2)
ggplot(df, aes(x = group, y = value)) +
    # 箱线图的颜色规则
    geom_boxplot(aes(color = group), fill = NA, outlier.shape = NA) +
    scale_color_manual(values = c("a" = "black", "b" = "blue", "c" = "brown")) +
    # 启动新的颜色刻度，后续图层使用独立规则
    new_scale_color() +
    # 散点边框的颜色规则
    geom_point(aes(color = group), shape = 21, 
               position = position_jitter(seed = 1, width = .05), 
               alpha = 0.1) +
    scale_color_manual(values = c("a" = "red", "b" = "green", "c" = "yellow"))

两种方案都能实现需求：分组a、b、c的箱线图边框分别为黑、蓝、棕，对应散点边框为红、绿、黄。

内容的提问来源于stack exchange，提问作者Patrick