power-analysis 0.1.5

This package allows you to perform power analysis computations

power-analysis 💪🔍

power-analysis is a Python package for performing power analysis and calculating sample sizes for statistical models. The package provides classes for defining statistical models, performing power analysis, calculating sample sizes for two-sample t-tests, and conducting power analysis on observational panel data using a clustered bootstrap method.

Installation 📥

You can install the power_analysis package using pip:

pip install power-analysis

Usage 🧑‍💻

Panel Data Power Analysis

To use the PanelBootstrap class for power analysis on observational panel data, you can follow these steps:

1. Import the required packages and classes:

import numpy as np
import pandas as pd
from power_analysis import PanelBootstrap

2. Load your panel data into a pandas DataFrame:

data = pd.read_csv('your_data.csv')

3. Create a PanelBootstrap object with the required parameters:

power_analysis = PanelBootstrap(data, outcome_var='outcome', treatment_var='treatment', individual_var='individual', random_seed=42)

4. Use the fit_model method to fit a linear regression model and obtain the treatment effect and p-value:

treatment_effect, p_value = power_analysis.fit_model(data)

5. Perform a clustered bootstrap analysis to obtain the mean effect size, standard deviation of effect sizes, and a list of p-values:

mean_effect_size, std_effect_size, p_values = power_analysis.clustered_bootstrap(n_bootstrap=1000)

6. Calculate the statistical power when varying the number of observations (N) or effect sizes:

n_values = [50, 100, 150, 200, 250, 300]
alpha = 0.05
n_bootstrap = 1000

power_by_n = power_analysis.calculate_power_by_n(n_values, alpha, n_bootstrap)
power_by_effect_size = power_analysis.calculate_power_by_effect_size(effect_sizes, alpha, n_bootstrap)

Contributing 🤝

Contributions to power-analysis are welcome! If you find a bug or would like to suggest a new feature, please open an issue on GitHub.

License 📜

power-analysis is licensed under the MIT license. See LICENSE for more information.