EnvCausal

A Causal Inference Framework for Environmental Data Analysis

This repository contains all the necessary data and custom Python scripts to reproduce the results in the paper "Machine Learning–Aided Causal Inference Framework for Environmental Data Analysis: A COVID-19 Case Study".

Jupyter Notebook examples will be uploaded in future releases.

Install

pip install EnvCausal

Script descriptions

clustering.py -- A PCA-k-means clustering pipeline.
causal_estimate.py -- A dowhy package wrapper, results will be written in a txt file¹.
recover_network.py -- A SAM model for Bayesian network recovery, adjacent matrix will be returned.
result_plot.py -- A couple of functions for result visualization.

¹Inspired by Dowhy example: Iterating over multiple refutation tests

Acknowledgements

Some functions in the the following packages were utlized in the software:
DoWhy
EconML
CausalML
CausalDiscoveryToolbox

Thanks for the developers' effort on those fascinating causal inference tools!