如何在R Studio中逐步整理数据集？求对应R代码

Hey there! Since I can't see the before/after dataset images you shared, I’ll walk you through a general step-by-step workflow for cleaning datasets in R Studio that covers most common data wrangling scenarios. You can easily tweak these steps based on your specific structure changes (like reshaping, splitting columns, or merging values)!

1. 加载核心数据处理包 #

We’ll use the tidyverse family of packages—it’s your one-stop shop for most data cleaning tasks:

# Install if you haven't already
install.packages("tidyverse")

# Load the package
library(tidyverse)

If you need to handle Excel files, add the readxl package with install.packages("readxl") and load it too.

2. 导入原始数据集 #

Import your data based on its file format:

# For CSV files
raw_data <- read_csv("your_raw_dataset.csv")

# For Excel files (using readxl)
# raw_data <- read_excel("your_raw_dataset.xlsx")

3. 先做数据探索（了解你的数据） #

Before diving into cleaning, get a sense of your data’s structure and issues:

# View first 6 rows
head(raw_data)

# Check variable types, sample values, and missing data
str(raw_data)

# Get summary stats for numeric variables
summary(raw_data)

# Count missing values per column
colSums(is.na(raw_data))

4. 处理缺失值 #

Choose a strategy based on your data and analysis needs:

# Option 1: Drop rows with any missing values (use sparingly!)
clean_data <- raw_data %>% drop_na()

# Option 2: Fill numeric columns with their mean
clean_data <- raw_data %>%
  mutate(across(where(is.numeric), ~replace_na(., mean(., na.rm = TRUE))))

# Option 3: Fill categorical columns with their mode (most frequent value)
get_mode <- function(x) {
  ux <- unique(x)
  ux[which.max(tabulate(match(x, ux)))]
}
clean_data <- raw_data %>%
  mutate(across(where(is.factor) | where(is.character), ~replace_na(., get_mode(.))))

5. 调整变量类型 #

Fix incorrectly assigned data types (e.g., dates stored as text, categories stored as numbers):

# Convert a character column to a factor (for categories)
clean_data <- clean_data %>%
  mutate(category_column = as.factor(category_column))

# Convert a text date to a proper date format (adjust the format string to match your data)
clean_data <- clean_data %>%
  mutate(date_column = as.Date(date_column, format = "%Y-%m-%d"))

6. 重命名变量（让列名更清晰） #

Rename confusing or inconsistent column names:

clean_data <- clean_data %>%
  rename(
    customer_id = old_cust_id,
    purchase_amount = amt_spent,
    order_date = date_of_purchase
  )

7. 筛选/过滤行 #

Keep only the rows relevant to your analysis:

# Example: Keep only rows where age is over 18 and purchase amount > 0
clean_data <- clean_data %>%
  filter(age > 18, purchase_amount > 0)

8. 重塑数据结构（宽转长/长转宽） #

If your before/after data involves switching between wide and long formats:

# Wide to long: Convert multiple columns into key-value pairs
long_data <- clean_data %>%
  pivot_longer(
    cols = starts_with("sales_"),  # Target columns to reshape
    names_to = "quarter",          # Name for the new key column
    values_to = "sales_amount"     # Name for the new value column
  )

# Long to wide: Reverse the above
wide_data <- long_data %>%
  pivot_wider(
    names_from = quarter,
    values_from = sales_amount
  )

9. 去除重复行 #

Remove duplicate entries that might skew your analysis:

clean_data <- clean_data %>% distinct()

10. 最终检查 #

Verify that your data is ready for analysis:

head(clean_data)
str(clean_data)
colSums(is.na(clean_data))  # Confirm no unexpected missing values

If you can share specific details about how your raw dataset differs from the desired output (e.g., "I need to split a full name column into first and last name" or "I want to merge two columns into one"), I can tailor the code to your exact needs!

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