使用ggplot2的facet_wrap自由刻度时调整Y轴范围以显示全部标签

问题背景 #

使用ggplot2绘制带facet_wrap(scales = "free_y")的柱状图时，geom_text添加的标签部分超出了绘图区域，需要在自动生成的Y轴范围基础上，给每个子图的Y轴上限增加一定比例，确保标签完整显示。

原代码 #

数据生成代码 #

# Load required libraries
library(dplyr)
library(tidyr)

# Function to generate random values
generate_random_values <- function(n_rows) {
  tibble(
    Category = rep(paste0("Cat_", seq_len(n_rows/12)), each = 12),
    budget = runif(n_rows, 1000, 150000),
    amount = runif(n_rows, -150000, 0),
    percent = runif(n_rows, 0, 100),
    month = rep(month.abb, n_rows/12),
  )
}

# Create a random data frame with similar structure
random_df <- generate_random_values(n_rows = 240)

绘图代码 #

ggplot(random_df, aes(y = abs(amount), x = month, fill = ifelse(percent > 0, "green", "red"))) + 
  geom_col(aes(type = "Actual spent")) + 
  geom_line(aes(y = budget, x = month, group = Category, color = "Budget", type = "Budget")) +
  facet_wrap(~Category, ncol = 4, scales = "free_y") +
  scale_fill_manual(name = "Type", values = c("green" = "green", "red" = "red"), 
                    labels = c("Actual spent (positive)", "Actual spent (negative)")) +
  scale_color_manual(name = "Type", values = c("Budget" = "black"), labels = "Budget") +
  theme_minimal() + 
  theme(axis.text.x = element_text(angle = 90)) + 
  geom_text(aes(y = floor(abs(amount)), label = floor(abs(amount))), size = 3, angle = 90, hjust = -0.1)

解决方案 #

由于facet_wrap(scales="free_y")会覆盖全局ylim设置，需要先提取自动计算的子图Y轴范围，再逐个调整：

修改后的完整代码 #

library(ggplot2)
library(grid)

# 1. 生成基础绘图对象
p <- ggplot(random_df, aes(y = abs(amount), x = month, fill = ifelse(percent > 0, "green", "red"))) + 
  geom_col(aes(type = "Actual spent")) + 
  geom_line(aes(y = budget, x = month, group = Category, color = "Budget", type = "Budget")) +
  facet_wrap(~Category, ncol = 4, scales = "free_y") +
  scale_fill_manual(name = "Type", values = c("green" = "green", "red" = "red"), 
                    labels = c("Actual spent (positive)", "Actual spent (negative)")) +
  scale_color_manual(name = "Type", values = c("Budget" = "black"), labels = "Budget") +
  theme_minimal() + 
  theme(axis.text.x = element_text(angle = 90)) + 
  geom_text(aes(y = floor(abs(amount)), label = floor(abs(amount))), size = 3, angle = 90, hjust = -0.1)

# 2. 提取绘图布局信息
pb <- ggplot_build(p)

# 3. 调整每个子图的Y轴上限（这里设置为原上限的1.1倍，可根据需求调整比例）
scale_factor <- 1.1
for(i in seq_along(pb$layout$panel_params)){
  original_y_max <- pb$layout$panel_params[[i]]$y$limits[2]
  pb$layout$panel_params[[i]]$y$limits[2] <- original_y_max * scale_factor
}

# 4. 转换为gtable并绘制
pg <- ggplot_gtable(pb)
grid.draw(pg)

关键说明 #

• 先通过ggplot_build()获取ggplot内部计算的面板参数，其中包含每个子图自动生成的Y轴范围
• 遍历每个面板，将Y轴上限乘以自定义的缩放因子（示例中为1.1，即增加10%）
• 最后用ggplot_gtable()转换为可绘制的gtable对象，再用grid.draw()输出图形

这样就能保证每个子图的Y轴范围在自动适配数据的基础上，预留足够空间显示标签，同时不破坏free_y的自适应特性。

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