ggplot2绘图添加颜色图例遇阻：调整aes后图例显示异常

我懂你现在的困扰——把颜色参数移到aes()里后图例虽然显示了，但颜色完全不对，这是因为我们需要明确给每个统计量指定对应的颜色映射，而不是让ggplot自动分配。下面给你两个解决方案，先看直接修改现有代码的方法：

方案1：修改现有图层的颜色映射 #

核心思路是给每个统计量（第一四分位数、均值、第三四分位数）定义一个分类标签，把这个标签放到aes(color = ...)里，再用scale_color_manual()把标签和你想要的颜色绑定起来：

p3 <- ggplot() +
  # 第一四分位数：指定分类标签为"1st Quartile"
  geom_point(data = dv, aes(x = year, y = X1st.Qu., group = 1, color = "1st Quartile"), 
             shape = 95, size = 8) +
  # 均值：指定分类标签为"Mean"（同时给点和线用同一个标签）
  geom_point(data = dv, aes(x = year, y = Mean, group = 1, color = "Mean"), 
             shape = 21, size = 2) +
  geom_line(data = dv, aes(x = year, y = Mean, group = 1, color = "Mean"), 
             size = .5) +
  # 第三四分位数：指定分类标签为"3rd Quartile"
  geom_point(data = dv, aes(x = year, y = X3rd.Qu., group = 1, color = "3rd Quartile"), 
             shape = 95, size = 8) +
  # 手动绑定标签和颜色，同时设置图例名称
  scale_color_manual(name = "Statistics",
                     values = c("1st Quartile" = "yellow",
                                "Mean" = "blue",
                                "3rd Quartile" = "red")) +
  # 合并theme设置，让代码更简洁
  theme(axis.text.x = element_text(angle = 90, size = 7),
        plot.title = element_text(hjust = 0.5)) +
  ggtitle("Statistical description per year") +
  labs(y = "Statistics", x = "Year")

这样修改后，图例就会正确显示每个统计量对应的颜色了——因为我们明确告诉了ggplot："1st Quartile"对应黄色，"Mean"对应蓝色，"3rd Quartile"对应红色。

方案2：整理数据为长格式（更推荐） #

如果后续要添加更多统计量，重复写多个图层会很麻烦，建议把宽格式的数据转成长格式，这样只用少量图层就能搞定：

首先用tidyr包转换数据：

library(tidyr)

dv_long <- dv %>%
  pivot_longer(cols = c(X1st.Qu., Mean, X3rd.Qu.),
               names_to = "stat_type",
               values_to = "value")

然后画图：

p3 <- ggplot(dv_long, aes(x = year, y = value, group = stat_type, color = stat_type)) +
  geom_point(aes(shape = stat_type, size = stat_type)) +
  geom_line(data = filter(dv_long, stat_type == "Mean"), size = .5) +
  # 指定颜色、形状、大小的映射
  scale_color_manual(name = "Statistics",
                     values = c("X1st.Qu." = "yellow",
                                "Mean" = "blue",
                                "X3rd.Qu." = "red")) +
  scale_shape_manual(name = "Statistics",
                     values = c("X1st.Qu." = 95,
                                "Mean" = 21,
                                "X3rd.Qu." = 95)) +
  scale_size_manual(name = "Statistics",
                    values = c("X1st.Qu." = 8,
                               "Mean" = 2,
                               "X3rd.Qu." = 8)) +
  theme(axis.text.x = element_text(angle = 90, size = 7),
        plot.title = element_text(hjust = 0.5)) +
  ggtitle("Statistical description per year") +
  labs(y = "Statistics", x = "Year")

这种方式代码更简洁，后续维护和扩展也更方便。

内容的提问来源于stack exchange，提问作者user9737581