如何使用ggplot2绘制发散式堆叠条形图？附可复现示例数据

Absolutely! You can definitely create a diverging stacked bar chart like the one you referenced using ggplot2. Below is a complete, reproducible example including sample data, quick data prep, and plotting code to match what you're looking for.

1. 准备环境与可复现示例数据 #

First, we’ll use the tidyverse package (which includes ggplot2 and handy data manipulation tools) and generate a consistent dataset with both positive and negative values—this is what creates the diverging "split" effect.

library(tidyverse)

# Set seed to ensure data is reproducible every time you run the code
set.seed(456)

# Create sample data: Main categories with sub-categories that have positive/negative contributions
df <- tibble(
  category = rep(c("Group 1", "Group 2", "Group 3", "Group 4"), each = 4),
  sub_category = rep(c("Factor A", "Factor B", "Factor C", "Factor D"), 4),
  value = c(
    # Group 1: 2 positive, 2 negative values
    runif(2, 10, 30), runif(2, -25, -5),
    # Group 2: 3 positive, 1 negative value
    runif(3, 5, 20), runif(1, -15, -5),
    # Group 3: 1 positive, 3 negative values
    runif(1, 15, 25), runif(3, -30, -10),
    # Group 4: 2 positive, 2 negative values
    runif(2, 20, 40), runif(2, -35, -15)
  )
)

2. 绘制水平发散堆叠条形图（贴合常见的"右侧"样式） #

This code will generate a horizontal diverging stacked bar chart, where positive values stack to the right of the 0 line and negative values stack to the left—this is likely what you’re referring to as the "right-side" style:

ggplot(df, aes(y = category, x = value, fill = sub_category)) +
  # Stack bars automatically: positive values go right, negative go left
  geom_col(position = "stack", width = 0.7) +
  # Add a bold reference line at 0 to clearly separate positive/negative
  geom_vline(xintercept = 0, color = "black", linewidth = 1) +
  # Use a color palette that makes sub-categories easy to tell apart
  scale_fill_brewer(palette = "Paired") +
  # Add clear labels and a centered title
  labs(
    y = "Category",
    x = "Value",
    fill = "Sub-category",
    title = "Horizontal Diverging Stacked Bar Chart"
  ) +
  # Clean up the theme for better readability
  theme_minimal() +
  theme(
    plot.title = element_text(hjust = 0.5, size = 14, face = "bold"),
    axis.title = element_text(size = 12),
    legend.position = "bottom"
  )

3. 绘制垂直版本（可选） #

If you prefer a vertical layout, just swap the x and y aesthetics and adjust the reference line:

ggplot(df, aes(x = category, y = value, fill = sub_category)) +
  geom_col(position = "stack", width = 0.7) +
  geom_hline(yintercept = 0, color = "black", linewidth = 1) +
  scale_fill_brewer(palette = "Paired") +
  labs(
    x = "Category",
    y = "Value",
    fill = "Sub-category",
    title = "Vertical Diverging Stacked Bar Chart"
  ) +
  theme_minimal() +
  theme(
    plot.title = element_text(hjust = 0.5, size = 14, face = "bold"),
    axis.title = element_text(size = 12),
    legend.position = "bottom"
  )

关键细节说明 #

• The magic happens with geom_col() and position = "stack": ggplot automatically handles stacking positive values in one direction and negative values in the opposite, creating the diverging effect from the 0 line.
• You can tweak the color palette, bar width, or theme elements to match your specific design needs.
• The set.seed() function ensures your sample data stays consistent every time you run the code, so your chart will look identical to the example.

内容的提问来源于stack exchange，提问作者rafa.pereira