如何用R绘制带须线、含实验室值与MPV的非箱线图？

1. 数据预处理：转成长格式 #

首先得把宽格式数据（一列实验室值、一列MPV值）转成长格式，新增一个类别变量（比如type）区分两种数值，这样才能在ggplot里统一映射样式。

示例代码：

library(tidyr)
# 假设原数据框df包含lab_value（实验室值）和mpv_value（MPV）列
df_long <- df %>%
  pivot_longer(
    cols = c(lab_value, mpv_value),
    names_to = "type",
    values_to = "result",
    names_transform = list(type = ~ifelse(.x == "lab_value", "实验室值", "MPV"))
  )

2. 核心散点图层构建 #

用geom_point()通过shape参数区分实心/空心点，搭配颜色映射实现需求：

• 实验室值：用实心圆点（shape=19），指定彩色
• MPV：用空心圆点（shape=21），填充白色保证空心效果

基础代码框架：

library(ggplot2)
ggplot(df_long, aes(x = type, y = result)) +
  # 绘制实心实验室值点
  geom_point(
    data = subset(df_long, type == "实验室值"),
    shape = 19, color = "#2ca02c", size = 3
  ) +
  # 绘制空心MPV点
  geom_point(
    data = subset(df_long, type == "MPV"),
    shape = 21, color = "#2ca02c", fill = "white", size = 3
  ) +
  # 美化标签与主题
  labs(x = "数值类型", y = "检测结果", title = "实验室值与MPV对比") +
  theme_minimal()

3. 进阶优化（按需选择） #

配对样本连线 #

如果是每个样本对应一组实验室值和MPV，用geom_line()连接同一样本的两个点，更直观展示差异：

# 假设数据包含sample_id（样本标识）列
ggplot(df_long, aes(x = type, y = result, group = sample_id)) +
  geom_line(color = "gray", alpha = 0.5) + # 灰色半透明连接线
  geom_point(data = subset(df_long, type == "实验室值"), shape = 19, color = "#2ca02c", size = 3) +
  geom_point(data = subset(df_long, type == "MPV"), shape = 21, color = "#2ca02c", fill = "white", size = 3) +
  labs(x = "数值类型", y = "检测结果", title = "配对样本：实验室值与MPV对比") +
  theme_minimal()

叠加分布背景 #

如果需要展示整体分布，可在散点下方叠加箱线图/小提琴图，用alpha调整透明度避免遮挡：

ggplot(df_long, aes(x = type, y = result)) +
  geom_boxplot(width = 0.3, alpha = 0.2) + # 半透明箱线图
  geom_point(data = subset(df_long, type == "实验室值"), shape = 19, color = "#2ca02c", size = 3, position = position_jitter(width = 0.1)) +
  geom_point(data = subset(df_long, type == "MPV"), shape = 21, color = "#2ca02c", fill = "white", size = 3, position = position_jitter(width = 0.1)) +
  labs(x = "数值类型", y = "检测结果", title = "实验室值与MPV分布对比") +
  theme_minimal()

关键注意点 #

• 长格式数据是ggplot灵活映射的基础，务必先完成格式转换
• ggplot形状参数：shape=19为实心圆，shape=21为可填充的空心圆，可根据喜好替换其他形状
• 配对数据必须指定group=样本标识，否则连接线会混乱

内容的提问来源于stack exchange，提问作者SharonP