SynthCtrl

A Python library for implementing Synthetic Control methods for causal inference, including the Classical Synthetic Control method and Synthetic Difference-in-Differences (SDID).

Overview

Synthetic Control is a statistical method used for comparative case studies. It constructs a weighted combination of control units to create a synthetic version of the treated unit, allowing for the estimation of causal effects in settings where a single unit receives treatment and multiple units remain untreated.

This library provides implementations of:

• Classical Synthetic Control (Abadie & Gardeazabal, 2003)
• Synthetic Difference-in-Differences (Arkhangelsky et al., 2019)

Installation

pip install SynthCtrl

Features

• Easy-to-use API with scikit-learn-like interfaces
• Bootstrap for statistical inference
• Comprehensive visualization tools

Quick Start

import pandas as pd
from synthetic_control import ClassicSyntheticControl

data = pd.read_csv("california_smoking.csv")

sc = ClassicSyntheticControl(
    data=data,
    metric="cigarettes",
    period_index="year",
    unit_id="state",
    treated="california",
    after_treatment="after_treatment"
)

sc.fit()

predictions = sc.predict()

effect = sc.estimate_effect()
print(f"Average Treatment Effect: {effect['att']:.4f}")

bootstrap_results = sc.bootstrap_effect()
print(f"Standard Error: {bootstrap_results['se']:.2f}")
print(f"95% CI: [{bootstrap_results['ci_lower']:.2f}, {bootstrap_results['ci_upper']:.2f}]")

sc.plot_model_results(figsize=(14, 7), show=True)

Using Synthetic Difference-in-Differences

from synthetic_control import SyntheticDIDModel

sdid_model = SyntheticDIDModel(
    data=data,
    metric="cigarettes",
    period_index="year", 
    unit_id="state",
    treated="california",
    after_treatment="after_treatment"
)

sdid_model.fit()

sdid_model.plot_model_results(figsize=(14, 7), show=True)

Documentation

For detailed documentation, visit the GitHub pages.

Examples

The examples/ directory contains Jupyter notebooks demonstrating various use cases:

• Basic usage with California smoking data
• Advanced features and customization
• Comparison of different methods

Citation

If you use this library in your research, please cite:

@software{SynthCtrl_python,
  author = {Yaroslav Rogoza},
  title = {SynthCtrl: A Python Library for Causal Inference},
  year = {2025},
  url = {https://github.com/123yaroslav/SynthCtrl},
}

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.