CausalEstimate 0.6.1

A library for estimates of causal effects.

CausalEstimate

---

CausalEstimate is a Python library designed for causal inference, providing a suite of methods to estimate treatment effects from observational data. It includes doubly robust techniques such as Targeted Maximum Likelihood Estimation (TMLE) alongside propensity score-based methods like inverse probability weighting (IPW) and matching. The library is built for flexibility and ease of use, integrating seamlessly with pandas and supporting bootstrap-based standard error estimation.

---

Features

• Causal inference methods including doubly robust estimators (TMLE, AIPW)
• Tools for estimating average treatment effects (ATE, ATT, RR...)
• Bootstrap standard error estimation

---

Installation

To install the required dependencies, run:

pip install CausalEstimate

For local installation, clone the repository and run:

pip install -e .

---

Usage

Example: Using the Estimator

Here's an example of how to use the Estimator class to compute effects:

import pandas as pd
import numpy as np
from CausalEstimate.interface.estimator import Estimator

# Simulate data
np.random.seed(42)
n = 1000
ps = np.random.uniform(0, 1, n)
treatment = np.random.binomial(1, ps)
outcome = 2 + 0.5*treatment + np.random.normal(0, 1, n)

df = pd.DataFrame({
    'treatment': treatment,
    'outcome': outcome,
    'ps': ps
})

# Create an Estimator object
estimator = Estimator(methods=['AIPW'], effect_type='ATE')

# Compute effects
results = estimator.compute_effect(
    df,
    treatment_col='treatment',
    outcome_col='outcome',
    ps_col='ps',
    bootstrap=True,
    n_bootstraps=100,
    method_args={},
    apply_common_support=False,
    common_support_threshold=0.1
)

print(results)

Example: Matching

Here is an example of how to use the matching functionality in a Jupyter notebook:

import numpy as np
import pandas as pd
from CausalEstimate.matching import match_optimal

# Simulate data
ps = np.array([0.3, 0.90, 0.5, 0.34, 0.351, 0.32, 0.35, 0.81, 0.79, 0.77, 0.90, 0.6, 0.52, 0.55])
treated = np.array([1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])
ids = np.array([101, 102, 103, 103, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211])

df = pd.DataFrame({
    'PID': ids,
    'treatment': treated,
    'ps': ps
})

# Perform matching
result = match_optimal(df, n_controls=3, caliper=0.1)
print(result)

This example demonstrates how to:

1. Create an Estimator object with a specified method (AIPW in this case)
2. Use the compute_effect method to estimate the Average Treatment Effect (ATE)
3. Apply bootstrap for standard error estimation

---

Development

Please refer to the CONTRIBUTING.md file for development guidelines.

---

Attributions

If you use CausalEstimate in your research, please cite it using the following BibTeX entry:

@software{causalestimate,
  author = {Kiril Klein, ...},
  title = {CausalEstimate: A Python Library for Causal Inference},
  year = {2024},
  url = {https://github.com/kirilklein/CausalEstimate},
  version = {X.Y.Z},
  note = {GitHub repository}
}

Contact

License

This project is licensed under the MIT License.