ggplot使用facet_grid时如何让geom_rect的透明度保持一致

问题成因 #

你当前的写法中geom_rect继承了全局的数据集映射，每个分面内有多少条数据，对应的矩形就会被重复绘制多少次：叠加的层数越多，实际显示的颜色就越深（透明度越低），不同分面的数据量不同，最终就出现了透明度不一致的问题。

解决方法 #

方法1：使用annotate绘制背景（推荐） #

annotate不会和数据集绑定，每个分面仅绘制一次矩形，透明度可以保持完全统一，修改后的代码如下：

ggplot(df, aes(x=Dem) )+
  theme_bw()+
  # 正区间背景
  annotate("rect", ymin = 0, ymax = 0.2, xmin = -Inf, xmax = Inf, fill = 'negligible', alpha = .2)+
  annotate("rect", ymin = 0.2, ymax = 0.5, xmin = -Inf, xmax = Inf, fill = 'small', alpha = .2)+
  annotate("rect", ymin = 0.5, ymax = 0.8, xmin = -Inf, xmax = Inf, fill = 'medium', alpha = .2)+
  annotate("rect", ymin = 0.8, ymax = Inf, xmin = -Inf, xmax = Inf, fill = 'large', alpha = .2)+
  # 负区间背景
  annotate("rect", ymin = 0, ymax = -0.2, xmin = -Inf, xmax = Inf, fill = 'negligible', alpha = .2)+
  annotate("rect", ymin = -0.2, ymax = -0.5, xmin = -Inf, xmax = Inf, fill = 'small', alpha = .2)+
  annotate("rect", ymin = -0.5, ymax = -0.8, xmin = -Inf, xmax = Inf, fill = 'medium', alpha = .2)+
  annotate("rect", ymin = -0.8, ymax = -Inf, xmin = -Inf, xmax = Inf, fill = 'large', alpha = .2)+
  # 原有数据图层
  geom_point(aes(y=effect), stat="identity")+
  geom_errorbar(aes(ymin=lower, ymax=upper), width=.2)+
  facet_grid(Subject~cat, scales="free")+
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

方法2：单独绑定空数据框保留图例 #

如果需要保留填充色的分类图例，可以单独创建一个空数据框给geom_rect使用，避免重复绘制：

# 创建空的背景映射数据框，不会触发重复绘制
bg_data <- data.frame()

ggplot(df, aes(x=Dem) )+
  theme_bw()+
  # 正区间背景，关闭全局美学继承
  geom_rect(data = bg_data, aes(ymin = 0, ymax = 0.2, xmin = -Inf, xmax = Inf, fill = 'negligible'), alpha = .2, inherit.aes = FALSE)+
  geom_rect(data = bg_data, aes(ymin = 0.2, ymax = 0.5, xmin = -Inf, xmax = Inf, fill = 'small'), alpha = .2, inherit.aes = FALSE)+
  geom_rect(data = bg_data, aes(ymin = 0.5, ymax = 0.8, xmin = -Inf, xmax = Inf, fill = 'medium'), alpha = .2, inherit.aes = FALSE)+
  geom_rect(data = bg_data, aes(ymin = 0.8, ymax = Inf, xmin = -Inf, xmax = Inf, fill = 'large'), alpha = .2, inherit.aes = FALSE)+
  # 负区间背景
  geom_rect(data = bg_data, aes(ymin = 0, ymax = -0.2, xmin = -Inf, xmax = Inf, fill = 'negligible'), alpha = .2, inherit.aes = FALSE)+
  geom_rect(data = bg_data, aes(ymin = -0.2, ymax = -0.5, xmin = -Inf, xmax = Inf, fill = 'small'), alpha = .2, inherit.aes = FALSE)+
  geom_rect(data = bg_data, aes(ymin = -0.5, ymax = -0.8, xmin = -Inf, xmax = Inf, fill = 'medium'), alpha = .2, inherit.aes = FALSE)+
  geom_rect(data = bg_data, aes(ymin = -0.8, ymax = -Inf, xmin = -Inf, xmax = Inf, fill = 'large'), alpha = .2, inherit.aes = FALSE)+
  # 原有数据图层
  geom_point(aes(y=effect), stat="identity")+
  geom_errorbar(aes(ymin=lower, ymax=upper), width=.2)+
  facet_grid(Subject~cat, scales="free")+
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

内容的提问来源于stack exchange，提问作者yuliaUU