为R数据集创建循环/函数，按多维度分区域计算结果占比

需求明确 #

需要针对数据集里的outcome1、outcome2、outcome3三个结果变量，分别按sex、age_group、income、education这四个维度计算各类别的占比；同时要按照geography字段的area1和area2生成两个独立的结果输出。

改进后的代码实现 #

library(tidyverse)

df <- data.frame (
  outcome1  = c("poor", "good", "excellent", "poor", "good", "poor", "poor", "excellent"),
  outcome2 = c("good", "excellent", "excellent", "poor", "excellent", "poor", "excellent", "poor"),
  outcome3 = c("poor", "poor", "excellent", "poor", "poor", "poor", "excellent", "good"),
  sex = c("F", "M", "M", "F", "F", "M", "F", "M"),
  age_group = c("50-54", "60-64", "80+", "70-74", "40-44", "45-49", "60-64", "65-69"),
  income = c("$<40,000", "$50,000-79,000", "$80,000-110,000", "$111,000+", "$<40,000", "$<40,000", "$50,000-79,000", "$80,000-110,000"),
  education = c("HS", "College", "Bachelors", "Masters", "HS", "College", "Bachelors", "Masters"),
  geography= c("area1", "area2", "area1", "area2", "area2", "area1", "area2", "area1")
)

# 定义占比计算函数
calc_proportion <- function(outcome_col, group_col, data) {
  data %>%
    group_by(geography, {{outcome_col}}, {{group_col}}) %>%
    summarise(count = n(), .groups = "drop_last") %>%
    mutate(total = sum(count), proportion = count / total * 100) %>%
    ungroup()
}

# 指定需要处理的变量组合
outcomes <- c("outcome1", "outcome2", "outcome3")
groups <- c("sex", "age_group", "income", "education")

# 批量计算所有组合的占比
all_results <- crossing(outcome = outcomes, group = groups) %>%
  mutate(result = map2(outcome, group, ~calc_proportion(.x, .y, df))) %>%
  unnest(result)

# 拆分出area1和area2的独立结果
area1_results <- all_results %>% filter(geography == "area1")
area2_results <- all_results %>% filter(geography == "area2")

# 查看结果示例
print(area1_results %>% head())
print(area2_results %>% head())

代码说明 #

• 函数复用：calc_proportion函数可适配任意结果变量和分组维度，自动按地区、结果、分组维度统计数量，计算每组占对应地区-结果组合的总比例，避免了硬编码总数的错误。
• 批量处理：用crossing生成所有结果变量与分组维度的组合，通过map2批量调用计算函数，一次性完成所有统计需求。
• 结果拆分：最后按geography筛选出两个地区的独立结果，方便后续分析或导出。

内容的提问来源于stack exchange，提问作者R_coder_new