如何使用tidyverse简化查询员工流失率最高与最低部门的代码

Hey there! Your original code logic is solid, but we can streamline it drastically using Tidyverse's chained operations and built-in functions—no need for multiple intermediate data frames. Let's break down the optimized version:

Full Simplified Code #

library(tidyverse)

# Calculate turnover per department and grab top/bottom results
turnover_summary <- df %>%
  group_by(department) %>%
  summarise(
    total_employees = n(),
    left_count = sum(left == "yes"),
    turnover_rate = left_count / total_employees
  ) %>%
  # Combine highest and lowest turnover rows
  bind_rows(
    slice_max(., turnover_rate, n = 1),
    slice_min(., turnover_rate, n = 1)
  ) %>%
  ungroup()

print(turnover_summary)

Key Optimizations Explained #

• group_by(department): Groups the data by department so all subsequent calculations are done per department subset.
• summarise(): Computes all necessary metrics in one step instead of splitting across data frames:
  • total_employees = n(): Counts total staff per department.
  • left_count = sum(left == "yes"): Counts employees who left (since TRUE converts to 1 and FALSE to 0, summing gives the count).
  • turnover_rate = left_count / total_employees: Calculates the turnover ratio directly.
• slice_max() + slice_min() + bind_rows(): Fetches the row with the highest turnover and the row with the lowest turnover, then combines them into a single result data frame. The . in slice_max(., ...) refers to the data coming down the pipe.
• ungroup(): Removes the department grouping to prevent unexpected behavior in any follow-up operations.

Even Shorter Version (Using a Quick Mean Trick) #

Since left == "yes" returns logical values (1 for TRUE, 0 for FALSE), taking the mean directly gives the turnover rate. This cuts down the summarise() step even more:

# Get top turnover department
top_turnover <- df %>%
  group_by(department) %>%
  summarise(turnover_rate = mean(left == "yes")) %>%
  slice_max(turnover_rate, n = 1)

# Get bottom turnover department
bottom_turnover <- df %>%
  group_by(department) %>%
  summarise(turnover_rate = mean(left == "yes")) %>%
  slice_min(turnover_rate, n = 1)

# Combine results
final_result <- bind_rows(top_turnover, bottom_turnover)

This keeps everything concise while maintaining readability—core Tidyverse principles!

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