使用 ggplot2 根据颜色为 geom_segment 添加第二个图例的最简洁方法答案

【问题标题】：Cleanest way to add second legend for geom_segment based on color using ggplot2使用 ggplot2 根据颜色为 geom_segment 添加第二个图例的最简洁方法
【发布时间】：2021-09-14 00:05:57
【问题描述】：

library(ggplot2)
library(scales)

df = data.frame(Type = rep(c('A', 'B'), 250), Index = seq(500), Value = cumsum(rnorm(500)))

colors = hue_pal()(3)
labels = c('Alpha', 'Beta', 'Gamma')

ggplot(df, aes(Index, Value)) +
    geom_line(aes(color = Type)) +
    geom_segment(x = 200, xend = 300, y = -8, yend = -8, color=colors[1]) +  # Label as "Alpha"
    geom_segment(x = 400, xend = 500, y = -4, yend = -4, color=colors[1]) +  # Label as "Alpha" 
    geom_segment(x =   0, xend = 100, y =  0, yend =  0, color=colors[2]) +  # Label as "Beta"
    geom_segment(x = 100, xend = 200, y =  4, yend =  4, color=colors[3]) +  # Label as "Gamma"
    geom_segment(x = 300, xend = 400, y =  8, yend =  8, color=colors[3])    # Label as "Gamma"

上面的代码产生下面的输出

我想添加第二个图例，标题为“分类”，条目“Alpha”“Beta”和“Gamma”对应于水平线段的三种颜色。 Adding a legend entry for geom_segment 的答案建议使用 scale_fill_manual 但它没有效果。我希望在 R 中有一种干净的方法来做到这一点。

【问题讨论】：

标签： r dataframe ggplot2 time-series data-visualization

【解决方案1】：

这不是特别“干净”，但我想不出更好的方法：

library(tidyverse)
df = data.frame(Type = rep(c('A', 'B'), 250), Index = seq(500), Value = cumsum(rnorm(500)))
scales::hue_pal()(3)
#> [1] "#F8766D" "#00BA38" "#619CFF"
seg_df <- data.frame(x = c(200, 400, 0, 100, 300),
                     xend = c(300, 500, 100, 200, 400),
                     y = c(-8, -4, 0, 4, 8), 
                     yend = c(-8, -4, 0, 4, 8), 
                     col = factor(c("#F8766D", "#F8766D", "#00BA38",
                                    "#619CFF", "#619CFF"),
                                  labels = c("Alpha", "Beta", "Gamma")))

ggplot(df, aes(Index, Value)) +
  geom_line(aes(col = Type), show.legend = TRUE) +
  geom_segment(data = seg_df, aes(x = x, xend = xend,
                                  y = y, yend = yend,
                                  fill = col),
               color = c("#F8766D", "#F8766D", "#00BA38",
                         "#619CFF", "#619CFF")) +
  scale_fill_manual("Classification",
                    values = c("Alpha", "Beta", "Gamma"),
                    guide = guide_legend(override.aes = list(
                      colour = c("#F8766D", "#00BA38", "#619CFF"))))

我很想看看其他人是否能想出一个更简单/更清洁的方法。

【讨论】：