修改组合的水平条形图（紧密设计）

Question

我有以下示例数据：

library(tidyverse)

df <- data.frame(col=rep(c("A_B", "A_C", "A_D", 
                           "B_A", "C_A", "D_A",
                           "B_C", "B_D", 
                           "C_B", "D_B",
                           "C_D", "D_C"), 2),
                 level=c(rep("lower_level", 12), rep("higher_level", 12)),
                 value=abs(rnorm(24, mean=5, sd=2)))%>% tibble()
df[c('origin', 'target')] <- str_split_fixed(df$col, '_', 2)
df <- df %>% select(c(origin, target, level, value))

我现在想为每个目标创建水平堆叠条形图 (df %>% filter(target=="A"))。我使用以下代码执行此操作：

# plot
p1 <- ggplot(data = df %>% filter(target=="A"), 
            aes(x = factor(level), y = value, fill = factor(origin)))+
  geom_bar(stat="identity", position="fill", width = .1) + 
  scale_fill_manual(values = c("A"="yellow", "B" = "green", "C"="red", "D"="blue")) +
  coord_flip()

由于我想稍后合并多个此类图（如下），因此我想

1. 删除 y 轴和条形之间的空白空间（或将其操纵为值 X)

2. 让 fill 标签显示在右侧

3. 左侧有一个值，表示“target: A”

4. 并且在所有绘图之间共享填充图例和 y 轴。

请参阅带注释的图：

作为参考，我使用此代码创建了其他绘图：

p2 <- ggplot(data = df %>% filter(target=="B"), 
       aes(x = factor(level), y = value, fill = factor(origin)))+
  geom_bar(stat="identity", position="fill", width = .1) +
  scale_fill_manual(values = c("A"="yellow", "B" = "green", "C"="red", "D"="blue")) +
  coord_flip()

p3 <- ggplot(data = df %>% filter(target=="C"), 
             aes(x = factor(level), y = value, fill = factor(origin)))+
  geom_bar(stat="identity", position="fill", width = .1) +
  scale_fill_manual(values = c("A"="yellow", "B" = "green", "C"="red", "D"="blue")) +
  coord_flip()

p4 <- ggplot(data = df %>% filter(target=="D"), 
             aes(x = factor(level), y = value, fill = factor(origin)))+
  geom_bar(stat="identity", position="fill", width = .1) +
  scale_fill_manual(values = c("A"="yellow", "B" = "green", "C"="red", "D"="blue")) +
  coord_flip()

并将它们与此代码结合起来（但如果需要，很乐意使用其他方式来组合它们）。

library("gridExtra")
grid.arrange(p1, p2, p3, p4, ncol = 1, nrow = 4)

Answer 1

听起来很像您只是想通过 target 进行分面。无需在这里拼接多个图。

ggplot(data = df %>% mutate(target = paste('Target:', target)), 
             aes(x = factor(level), y = value, fill = factor(origin)))+
  geom_col(position = "fill", width = 0.9) + 
  scale_fill_manual(values = c("A"="yellow", "B" = "green", 
                               "C"="red", "D"="blue"), name = 'origin') +
  facet_grid(target~., switch = 'y') +
  coord_flip() +
  theme(strip.placement = 'outside',
        strip.background = element_blank(),
        axis.title.y = element_blank())

Answer 2

两个建议_

1. 要消除轴和条之间的偏移，请将轴扩展设置为零
   scale_x_continuous(..., Expand = c(0,0))
2. 不要对数据框进行繁琐的子集化，而是使用 facet_wrap 或 facet_grid 选项ggplot 的：

ggplot(data = df, 
             aes(x = factor(level), y = value, fill = factor(origin))) +
       ## other plot instructions
       facet_wrap( ~target)

请参阅 ?facet_wrap 了解各种布局选项，例如绘图列数
3.无论如何，条形之间的垂直间距将根据输出尺寸进行调整（此处：图形高度）