R 分别总结许多分类变量

Question

library(dplyr)
data = data.frame(
  group = c("A", "A", "A", "B","B"),
  category1 = c("X", "Y", "Z", "Z", "X"),
  category2 = c("M", "Q", "M", "M", "L")
)

data %>% group_by(group, category2) %>% summarize(count = n()) %>% ungroup() %>%
  group_by(group) %>%
  mutate(percent = count/sum(count)*100)

data %>% group_by(group, category1) %>% summarize(count = n()) %>% ungroup() %>%
  group_by(group) %>%
  mutate(percent = count/sum(count)*100)

我想总结一下类别1和类别2。我可以像所示的那样单独进行操作，但我不知道如何将它们组合起来并获得相同的结果。 我尝试了这个，但它不正确：

data %>% group_by(group, category1, category2) %>% summarize(count = n()) %>% ungroup() %>%
  group_by(group) %>%
  mutate(percent = count/sum(count)*100)

期望的输出将是一个表格，其中包含独立于其他类别变量的组中每个类别变量的频率和百分比...

Answer 1

data <- data.frame(
  group = c("A", "A", "A", "B", "B"),
  category1 = c("X", "Y", "Z", "Z", "X"),
  category2 = c("M", "Q", "M", "M", "L")
)

library(tidyverse)

to_do_list <- c(
  "category2",
  "category1"
)

names(to_do_list) <- to_do_list

(result_df <- map(to_do_list, \(x){
  data |>
    group_by(group,
      category_value = !!sym(x)
    ) |>
    summarize(count = n()) |>
    ungroup() |>
    group_by(group) |>
    mutate(percent = count / sum(count) * 100)
}) |> list_rbind(names_to = "category_source"))

# A tibble: 9 × 5
  category_source group category_value count percent
  <chr>           <chr> <chr>          <int>   <dbl>
1 category2       A     M                  2    66.7
2 category2       A     Q                  1    33.3
3 category2       B     L                  1    50  
4 category2       B     M                  1    50  
5 category1       A     X                  1    33.3
6 category1       A     Y                  1    33.3
7 category1       A     Z                  1    33.3
8 category1       B     X                  1    50  
9 category1       B     Z                  1    50