Examples

The examples directory contains runnable Jupyter notebooks demonstrating mini-causal features. These notebooks provide practical, hands-on examples of how to use the library for various causal analysis tasks.

Available Notebooks

Causal Models (Classifier)

File: examples/causal_models_classifier_example.ipynb

This notebook demonstrates how to compare two classification models using the causality module. Learn how to:

Initialize and configure CausalModelsClassifier
Run causal comparison between models
Interpret the results
Visualize model differences

Key concepts covered:

Causal metrics for model comparison
Statistical tests for causal effects
Practical workflow for model evaluation

Causal Counterfactual

File: examples/mini_causal_causal_counterfactual_example.ipynb

This notebook demonstrates the complete counterfactual workflow using the causal_counter module. Learn how to:

Set up counterfactual analysis
Generate counterfactual explanations
Evaluate counterfactual quality
Interpret the results

Key concepts covered:

Counterfactual generation algorithms
Feature importance through counterfactuals
Model interpretability

Partial Counterfactual (Prostate Dataset)

File: examples/mini_causal_prostate_with_partial_counterfactual.ipynb

This notebook demonstrates targeted counterfactual analysis on the prostate dataset using the partial_counter module. Learn how to:

Focus on specific feature subsets
Run partial counterfactual analysis
Compare results with full counterfactuals
Interpret targeted interventions

Key concepts covered:

Partial counterfactual generation
Feature subset selection
Targeted causal analysis
Domain-specific application

Running the Examples

Prerequisites

Ensure you have Jupyter installed:

pip install jupyter

Optional packages for visualization:

pip install matplotlib seaborn

Launching Jupyter

Navigate to the repository root:
```
cd mini-causal
```
Start Jupyter Lab or Notebook:
```
jupyter lab
# or
jupyter notebook
```
Open the notebooks from the examples directory

Tips for Working with Examples

Start with the classifier example if you’re new to causal analysis
Each notebook includes detailed explanations and code comments
Modify the code to use your own datasets and models
Refer to the module documentation for API details

Note: Some examples may require additional datasets. These are typically included in the datasets directory or will be downloaded automatically.