Introduction

mini-causal is a lightweight Python library designed for causal analysis, model comparison, and counterfactual estimation. Built with simplicity and practicality in mind, it offers an accessible entry point for exploring causal inference without the complexity of larger frameworks.

What is Causal Analysis?

Causal analysis goes beyond traditional statistical methods to understand the cause-and-effect relationships between variables. Unlike correlation, which identifies associations, causal analysis seeks to determine whether one variable truly influences another.

Use Cases

• Feature importance analysis: Measure how each feature impacts model predictions

• Decision Support for Feature Selection: Identify which features have the most causal impact for better feature selection

• Measuring the impact of features on the model’s metrics: Quantify how features affect model performance metrics

The library is designed to fill in the gap for quantifying for features on the model’s metrics as well as for production workflow.