How to Create a Before-and-After Comparison Table for Cleaned Data (Google Colab)

Learn how to create a before-and-after comparison table in Python using pandas to clearly show the impact of your data cleaning steps.

Step 1: Set Up Google Colab and Load Data

import pandas as pd
from google.colab import files

uploaded = files.upload()
file_name = list(uploaded.keys())[0]

df = pd.read_csv(file_name)
df_original = df.copy()  # Preserve original data

Step 2: Apply Your Cleaning Steps

Example cleaning:

def clean_data(df):
    df = df.dropna(subset=['country'])
    
    for col in df.select_dtypes(include='number').columns:
        df[col] = df[col].fillna(df[col].median())
    
    for col in df.select_dtypes(include='object').columns:
        df[col] = df[col].str.strip().str.lower()
    
    return df

df_clean = clean_data(df)

Step 3: Create a Summary Comparison Table

Compare key metrics before vs after cleaning.

comparison = pd.DataFrame({
    "Metric": [
        "Row Count",
        "Column Count",
        "Missing Values",
        "Duplicate Rows"
    ],
    "Before Cleaning": [
        df_original.shape[0],
        df_original.shape[1],
        df_original.isna().sum().sum(),
        df_original.duplicated().sum()
    ],
    "After Cleaning": [
        df_clean.shape[0],
        df_clean.shape[1],
        df_clean.isna().sum().sum(),
        df_clean.duplicated().sum()
    ]
})

comparison

Step 4: Compare Specific Columns (Optional)

Track changes in important columns.

column_comparison = pd.DataFrame({
    "Original": df_original['Country Name'].head(10),
    "Cleaned": df_clean['Country Name'].head(10)
})

column_comparison

Step 5: Highlight Changes (Optional)

Show only rows that changed.

changes = df_original.compare(df_clean)
changes.head()

Step 6: Export the Comparison

comparison.to_csv("comparison_summary.csv", index=False)
files.download("comparison_summary.csv")

Key Rules

• Always keep a copy of original data (df_original)

• Compare structure (shape), quality (missing values), and duplicates

• Use .compare() for row-level differences

• Focus on the metrics stakeholders understand easily

What You Just Achieved

• Clear visibility of cleaning impact

• Reusable comparison framework

• Stakeholder-ready reporting

This approach makes your data cleaning measurable, not just assumed.

Advance Your Career With 16 Python Projects in Data & ML — All for $288.