Box-Cox Transformation Explorer

1. Enter positive numeric data (Box-Cox requires positive values)
2. Or generate sample skewed data for demonstration
3. Click "Find Optimal Lambda" to calculate the best transformation
4. Use the slider to manually explore different lambda values
5. Compare before/after distributions and Q-Q plots
6. Use common lambda buttons for standard transformations

• λ = 2: Square transformation (y²)
• λ = 1: No transformation (y)
• λ = 0.5: Square root transformation (√y)
• λ = 0: Natural log transformation (ln(y))
• λ = -1: Inverse transformation (1/y)
• Other values: Power transformations (y^λ)

The Box-Cox transformation is defined as:

• If λ ≠ 0: y(λ) = (y^λ - 1) / λ
• If λ = 0: y(λ) = ln(y)

The optimal λ is found by maximizing the log-likelihood function.

• Data is positively skewed (long right tail)
• Need to meet normality assumptions for statistical tests
• Want to stabilize variance across groups
• Preparing data for linear regression
• All values must be positive (add constant if needed)