Regression Residual Diagnostics Studio

Regression Core LINE Assumption Audit

Inspect the 4 essential diagnostic plots that determine whether an Ordinary Least Squares (OLS) regression model can be trusted for managerial decisions.

⚠️ "Break the Model" Interactive Fault Injectors

Toggle to see how assumption violations distort each diagnostic plot in real time.

Inject Quadratic Curvature

Simulates diminishing returns (violates Linearity)

Inject Funnel Heteroscedasticity

Expands variance with scale (violates Equal Variance)

Inject High-Leverage Outlier

Spikes Cook's D and tilts regression slope

Case Study Scenario

Preset Case:

Linearity

Evaluating...

Residuals should be evenly scattered around zero without systematic curvature.

Independence

Evaluating...

Durbin-Watson statistic checks for sequential correlation among errors.

Normality

Evaluating...

Residual quantiles should hug the 45-degree diagonal reference line.

Equal Variance

Evaluating...

Residual spread should remain constant across the fitted range (homoscedasticity).

1. Residuals vs. Fitted (Linearity)

Look for zero-line symmetry

2. Normal Q-Q Plot (Normality)

Points should track red line

3. Scale-Location (Equal Variance)

Look for flat horizontal spread

4. Cook's Distance (Influential Points)

Warning > 0.5, Critical > 1.0

Regression Residual Diagnostics Studio

⚠️ "Break the Model" Interactive Fault Injectors

Case Study Scenario

1. Residuals vs. Fitted (Linearity)

2. Normal Q-Q Plot (Normality)

3. Scale-Location (Equal Variance)

4. Cook's Distance (Influential Points)

Statistical & Managerial Report

Diagnostic Audit Summary:

Managerial Translation & Action Plan:

Python Statsmodels Diagnostics Snippet