如何在R中绘制堆叠条形图？团队特征数据可视化需求

1. 堆叠条形图（绝对值） #

要实现每支团队的主条形堆叠三个特征数值，首先需要将数据转换为长格式（tidy data），这是ggplot2的最优数据格式。

数据转换与绘图代码 #

# 加载所需包
library(ggplot2)
library(dplyr)
library(tidyr)

# 原始数据
Team <- c("A", "B", "C", "D", "E", "F", "G", "H", "I", "J", "K", "L", "M")
total <- c(557, 3, 3116, 201, 167, 1348, 877, 1444, 2, 1003, 25, 732, 2532)
characteristic1 <- c(111, 0, 566, 45, 77, 600, 356, 300, 0, 402, 3, 278, 312)
characteristic2 <- c(20, 0, 231, 14, 15, 27, 30, 78, 0, 48, 0, 29, 111)
characteristic3 <- c(1, 0, 29, 1, 0, 10, 11, 3, 0, 2, 9, 1, 3)

df <- data.frame(Team, total, characteristic1, characteristic2, characteristic3)

# 转换为长格式，提取特征列
df_long <- df %>%
  select(Team, characteristic1, characteristic2, characteristic3) %>%
  pivot_longer(cols = -Team, names_to = "Characteristic", values_to = "Value")

# 绘制堆叠条形图
ggplot(df_long, aes(x = Team, y = Value, fill = Characteristic)) +
  geom_bar(stat = "identity") +
  labs(title = "各团队特征数值堆叠条形图",
       x = "团队",
       y = "数值",
       fill = "特征类型") +
  theme_minimal()

如果需要对比特征数值与总人数，可以添加参考线和标注：

ggplot(df_long, aes(x = Team, y = Value, fill = Characteristic)) +
  geom_bar(stat = "identity") +
  # 添加总人数标注
  geom_text(data = df, aes(x = Team, y = total, label = total), vjust = -0.5, size = 3) +
  # 添加总人数虚线参考线
  geom_hline(data = df, aes(yintercept = total), linetype = "dashed", color = "gray50") +
  labs(title = "各团队特征数值与总人数对比",
       x = "团队",
       y = "数值",
       fill = "特征类型") +
  theme_minimal()

2. 堆叠条形图（百分比） #

要转换为占总人数的百分比，只需在数据预处理阶段计算特征占比即可：

数据转换与绘图代码 #

# 计算特征占总人数的百分比，处理NA值（总人数为0的团队）
df_percent <- df %>%
  mutate(
    pct1 = characteristic1 / total * 100,
    pct2 = characteristic2 / total * 100,
    pct3 = characteristic3 / total * 100
  ) %>%
  select(Team, pct1, pct2, pct3) %>%
  pivot_longer(cols = -Team, names_to = "Characteristic", values_to = "Percentage") %>%
  mutate(
    Characteristic = recode(Characteristic,
                           "pct1" = "特征1",
                           "pct2" = "特征2",
                           "pct3" = "特征3"),
    Percentage = replace_na(Percentage, 0)
  )

# 绘制百分比堆叠条形图
ggplot(df_percent, aes(x = Team, y = Percentage, fill = Characteristic)) +
  geom_bar(stat = "identity") +
  labs(title = "各团队特征占总人数百分比堆叠图",
       x = "团队",
       y = "占比 (%)",
       fill = "特征类型") +
  theme_minimal() +
  scale_y_continuous(limits = c(0, 100)) # 强制Y轴范围为0-100%

3. 非条形图替代方案 #

雷达图 #

适合对比多维度特征在不同团队间的分布，尤其适配百分比数据：

library(fmsb)

# 整理雷达图所需格式：添加最大值、最小值行
df_radar <- df_percent %>%
  pivot_wider(names_from = Team, values_from = Percentage)

radar_data <- rbind(rep(100, 13), rep(0, 13), df_radar %>% select(-Characteristic))
rownames(radar_data) <- c("Max", "Min", df_radar$Characteristic)

# 绘制雷达图
radarchart(radar_data, 
           pcol = c("#00AFBB", "#E7B800", "#FC4E07"),
           pfcol = scales::alpha(c("#00AFBB", "#E7B800", "#FC4E07"), 0.3),
           plwd = 2,
           cglcol = "gray",
           cglty = 1,
           axislabcol = "gray",
           title = "各团队特征占比雷达图")
legend(x = 1.3, y = 1, legend = rownames(radar_data)[3:5], 
       col = c("#00AFBB", "#E7B800", "#FC4E07"), lty = 1, lwd = 2)

热力图 #

清晰展示不同团队与特征间的占比差异，适合快速定位极值：

ggplot(df_percent, aes(x = Team, y = Characteristic, fill = Percentage)) +
  geom_tile(color = "white") +
  geom_text(aes(label = sprintf("%.1f%%", Percentage)), color = "black", size = 3) +
  scale_fill_viridis_c(option = "plasma") +
  labs(title = "各团队特征占比热力图",
       x = "团队",
       y = "特征类型",
       fill = "占比 (%)") +
  theme_minimal()

散点图矩阵 #

探索总人数与各特征、特征之间的相关性：

library(GGally)

ggpairs(df, columns = 2:5, aes(color = Team)) +
  labs(title = "团队总人数与特征数值散点图矩阵") +
  theme_minimal()

内容的提问来源于stack exchange，提问作者marty