Raw data is often in formats that models or analysis tools cannot use directly.

Transformation = converting data into a usable form.

🔄 Common Transformations:

1. Type Conversion

  • "25" → 25 (string to integer)

2. Date Formatting

  • Convert "12/01/2024" → standard format like YYYY-MM-DD

3. Scaling Values

Important for machine learning models.

Why?
Different scales can bias models.

Example:

  • Age: 20–60
  • Salary: 20,000–1,00,000

Salary dominates unless scaled.

📏 Common Scaling Methods:

  • Normalization (0 to 1 range)
  • Standardization (mean = 0, std = 1)

💡 Key Idea:

Transformation ensures:

  • Consistency
  • Compatibility
  • Better model performance
PreviousNext