如何在R中绘制多变量多分组并列条形图以对比组间均值

1. 加载所需工具包 #

我们使用tidyverse套件完成数据清洗与可视化，先加载包：

# 未安装先执行 install.packages("tidyverse")
library(tidyverse)

2. 数据预处理 #

首先构造你的样例数据，再按需求计算分组均值：

# 构造样例数据框
df <- data.frame(
  Group = rep(c("Anxiety", "Control"), each = 5),
  Gender = c(0,1,0,1,1,0,0,1,1,0),
  Age = c(25,48,32,24,18,45,44,26,38,18),
  Anxiety_score = c(32,34,48,43,44,12,11,21,12,13),
  Depression_score = c(29,28,32,26,26,11,11,10,NA,10),
  IUS = c(12,11,12,12,15,3,5,5,2,1),
  OBSC = c(rep(NA,5), rep(NA,5)) # 你样例里OBSC未填值，实际使用替换为真实数据即可
)

# 计算分组均值：Gender的均值即为性别为1的占比，符合你的需求
df_mean <- df %>%
  group_by(Group) %>%
  summarise(across(where(is.numeric), ~mean(.x, na.rm = TRUE))) %>%
  # 转换为ggplot所需的长格式
  pivot_longer(cols = -Group, names_to = "variable", values_to = "mean_val")

3. 绘制并列条形图 #

满足Y轴放变量、两组条形相邻的要求，使用水平条形图实现：

ggplot(df_mean, aes(y = variable, x = mean_val, fill = Group)) +
  # position_dodge实现并列非堆叠
  geom_col(position = position_dodge(width = 0.8), width = 0.7) +
  # 坐标轴与图例标注，可按需修改
  labs(
    y = "变量",
    x = "组均值/性别占比",
    fill = "分组"
  ) +
  theme_bw()

可选美化调整 #

如果需要在条形上添加数值标签，可在ggplot代码中加入：

geom_text(aes(label = round(mean_val, 2)), position = position_dodge(width = 0.8), hjust = -0.1)

如果要调整分组配色，可加上：

scale_fill_manual(values = c("Anxiety" = "#f8766d", "Control" = "#00bfc4"))

内容的提问来源于stack exchange，提问作者user16518207