r - 在两个 ggplot 直方图上显示平均值和中位数

Question

我是新的 stackoverflow 用户，目前无法对原始帖子发表评论以提出问题。我遇到了以前的 stackoverflow 答案（https://stackoverflow.com/a/34045068/11799491），我想知道如何在此图中添加两条垂直线（组的平均值和组的中位数）。

我的尝试：我不知道如何在组变量“type”中添加

geom_vline(aes(xintercept = mean(diff), ), color="black") + 
geom_vline(aes(xintercept = median(diff), ), color="red") 

Answer 1

有几种不同的方法可以做到这一点，但我喜欢创建一个单独的汇总数据框，然后将其传递给 geom_vline 调用。这使您可以分析结果并轻松添加按类型自动排序和着色的多行：

library(tidyverse) 

df <-
  tibble(
    x = rnorm(40),
    category = rep(c(0, 1), each = 20)
  )

df_stats <-
  df %>% 
  group_by(category) %>% 
  summarize(
    mean = mean(x), 
    median = median(x)
  ) %>% 
  gather(key = key, value = value, mean:median)

df %>% 
  ggplot(aes(x = x)) +
  geom_histogram(bins = 20) +
  facet_wrap(~ category) +
  geom_vline(data = df_stats, aes(xintercept = value, color = key))

Answer 2

最简单的方法是按组预先计算均值和中位数type。我会用aggregate.

agg <- aggregate(diff ~ type, data, function(x) {
  c(mean = mean(x), median = median(x))
})
agg <- cbind(agg[1], agg[[2]])
agg <- reshape2::melt(agg, id.vars = "type")

library(ggplot2)

ggplot(data, aes(x = diff)) +
  geom_histogram() +
  geom_vline(data = agg, mapping = aes(xintercept = value,
                                       color = variable)) +
  facet_grid(~type) +
  theme_bw()