ggplot_比例表现02

单变量比例分布

条形图、堆叠条形图、并排条形图可以用来表现比例以及分组内的比例。饼图经常用来表现单个分组方式的各组比例。

ggplot2没有专门的饼图函数，可以用geom_col()与coord_polar()配合

R 复制代码

library(ggplot2)
library(ggridges)
library(dplyr)
library(GGally)

iris = read.csv('../../seaborn-data/iris.csv')
tips = read.csv('../../seaborn-data/tips.csv')

复制代码

Attaching package: 'dplyr'


The following objects are masked from 'package:stats':

    filter, lag


The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union

R 复制代码

tips_day_count <- tips |>
  select(day) |>
  count(day) |>
  mutate(ratio = n)

p <- ggplot(data = tips_day_count,mapping = aes(x = 1,y=ratio,fill=day))

p + geom_col() + coord_polar(theta = "y") +
  scale_x_continuous(name = NULL, breaks = NULL) +
  scale_y_continuous(name = NULL, breaks = NULL) +
  labs(fill = "Cut")

当然也可以用R自带的pie函数绘制饼图

R 复制代码

piepercent = paste(round(100*tips_day_count$n/sum(tips_day_count$n)), "%")
pie(tips_day_count$n,labels=piepercent,col=rainbow(12))
legend("topright", tips_day_count$day, cex=0.8, fill=rainbow(12))

再粗糙点，直接用geom_col(postion='stack)，也能达到饼状图的表达效果

R 复制代码

p + geom_col(position = 'stack')

有时需要比较不同组的某个变量的比例分布。

R 复制代码

ggplot(data = tips,mapping = aes(x = day,y=total_bill,fill=sex))+geom_col(position = 'dodge')

R 复制代码

ggplot(data = tips,mapping = aes(x = total_bill,fill=day))+geom_density(position = 'fill')

R 复制代码

ggplot(data = tips,mapping = aes(x = total_bill,fill=day))+geom_density_line(alpha=0.5,mapping = aes(y=..count..))

嵌套比例分布的比较

马赛克图是另外一种表现列联表频数的图形，用色块面积表示每个交叉类的频数。与堆叠条形图相比，这里用了条形宽度而不是条形高度表示大类频数。

R 复制代码

tips_day_count <- tips |>
  count(day,smoker)|>
  group_by(day) |>
  mutate(ratio = sum(n))

ggplot(tips_day_count,aes(y=n,x=day,width=ratio,fill=smoker))+
  geom_col(position = 'fill')+
  facet_grid(~day,scales = "free_x",space = "free_x")

R 复制代码

tips |>
  count(day,smoker)|>
  group_by(day) |>
  mutate(ratio = sum(n))

| day | smoker | n | ratio |

<chr>	<chr>	<int>	<int>
Fri	No	4	19
Fri	Yes	15	19
Sat	No	45	87
Sat	Yes	42	87
Sun	No	57	76
Sun	Yes	19	76
Thur	No	45	62
Thur	Yes	17	62
[A grouped_df: 8 × 4]