r - 如何绘制带有嵌套类别轴的图表?

标签 r ggplot2

我正在绘制具有类别和子类别的数据(请参见下面的示例数据),并且我想将它们嵌套显示(此示例是在 Excel 中创建的):

Nested categories from Excel
我在 R 中想到的最好的方法是创建一个具有所需名称的新列,如下所示:

df <- data.frame(main.cat = c("A", "A", "B", "B", "B", "C"),
                 second.cat = c("a1", "a2", "b1", "b2", "b3", "c1"),
                 value = c(2, 3, 4, 2.5, 1.5, 2.3))

df$x.labels <- paste(df$second.cat, df$main.cat, sep = "\n")

ggplot(data = df, aes(x = x.labels, y = value)) + geom_point()

这至少保留了两个级别的类别,但复制了所有主要类别标签:
enter image description here

有没有人知道更好的东西,它看起来更像 Excel 的输出?

最佳答案

我认为分面方法很好:

library(ggplot2)
library(gtable)
library(grid)

df <- data.frame(main.cat = c("A", "A", "B", "B", "B", "C"),
                 second.cat = c("a1", "a2", "b1", "b2", "b3", "c1"),
                 value = c(2, 3, 4, 2.5, 1.5, 2.3))

p = ggplot(data = df, aes(x = second.cat, y = value)) + 
   geom_point() + facet_grid(.~main.cat, space = "free_x", scales = "free_x") +
   theme(strip.background = element_rect(fill = NA))

但是,如果您想要更接近 excel 外观的东西,一种方法是使用 gtable函数提取 strip 并将其插入刻度线标签下方,然后在主要类别之间插入一些边界线。请注意,下面的代码特定于您的示例数据。
p = p + theme(panel.spacing = unit(0, "lines"))
g = ggplotGrob(p)
 gtable_show_layout(g)  # to see the layout

# Add a row below the x-axis tick mark labels,
# the same height as the strip
g = gtable_add_rows(g, g$height[7], 9)

# Get the strip grob
stripGrob = gtable_filter(g, "strip")

# Insert the strip grob into the new row
g = gtable_add_grob(g, stripGrob, 10, 5, 10, 9)

# remove the old strip
g = g[-7, ]

# Insert line grobs as boundary lines between major categories
linesGrob = linesGrob(gp = gpar(col = "grey75"))
for(i in c(6,8))  g = gtable_add_grob(g, linesGrob, t=8, l=i, b=9, r=i)

# Insert new columns of zero width to take the line grobs for the first and last boundary lines
for(i in c(4, 10)) {
   g = gtable_add_cols(g, unit(0, "lines"), i)
   g = gtable_add_grob(g, linesGrob, t=8, l=i+1, b=9, r=i+1)
}

grid.newpage()
grid.draw(g)

enter image description here

编辑 概括的粗略尝试
library(ggplot2)
library(gtable)
library(grid)

df <- data.frame(main.cat = c("A", "A", "B", "B", "C", "D"),
                 second.cat = c("a1", "a2", "b1", "b2", "c1", "d1"),
                 value = c(2, 3, 4, 2.5, 1.5, 2.3))

p = ggplot(data = df, aes(x = second.cat, y = value)) + 
   geom_point() + facet_grid(.~main.cat, space = "free_x", scales = "free_x") +
   theme(strip.background = element_rect(fill = NA))


p = p + theme(panel.spacing = unit(0, "lines"))
g = ggplotGrob(p)
 gtable_show_layout(g)  # to see the layout

# Get the indices for the panels (t=top, l=left, ...
panels <- c(subset(g$layout, grepl("panel", g$layout$name), se=t:r))

# Get the strip grob
stripGrob = gtable_filter(g, "strip")

 # Its height is
 height = stripGrob$height

# Add a row below the x-axis tick mark labels,
# the same height as the strip. 
g = gtable_add_rows(g, height, unique(panels$b+1))

# Insert the strip grob into the new row
g = gtable_add_grob(g, stripGrob, 
                      t = unique(panels$b+2), 
                      l = min(panels$l), 
                      r = max(panels$r))

# Insert line grobs as boundary lines between major categories
linesGrob = linesGrob(gp = gpar(col = "grey75"))
panelsR = panels$r[-length(panels$r)]
for(i in panelsR+1)  g = gtable_add_grob(g, linesGrob, 
                       t=unique(panels$b+1), 
                       l=i, 
                       b=unique(panels$b+2))

# Insert new columns of zero width to take the line grobs for the first and last boundary lines
 panelBound = c(4, max(panels$r)+1)
for(i in panelBound) {
   g = gtable_add_cols(g, unit(0, "lines"), i)
   g = gtable_add_grob(g, linesGrob, 
                    t=unique(panels$b+1), 
                    l=i+1, 
                    b=unique(panels$b+2))
}

# remove the old strip
g = g[-7, ]

# Draw it
grid.newpage()
grid.draw(g)

关于r - 如何绘制带有嵌套类别轴的图表?,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/28853786/

相关文章:

r - 使用 data.table 加速 rollapply

r - 使用 ggplot 或 R 中的任何其他方法根据计数绘制线宽(大小)

r - 更改密度直方图的 binwidth 以使概率总和为 1

r - 将自定义数字添加到 ggplot 图例

r - R中的圆形堆积条形图

r - sapply() 返回列表而不是向量

sql - 将 data.table 更新到 SQL 数据库?

r - 以 polylineDecorator 为例的传单插件和传单代理

r - R 中宽限月的 Emi 计算

r - 用图案填充地理空间多边形 - R