causers Documentation

High-performance statistical operations for Polars DataFrames, powered by Rust.

Contents:

Quick Start

Install causers from PyPI:

pip install causers

Basic usage:

import polars as pl
import causers

df = pl.DataFrame({"x": [1, 2, 3, 4, 5], "y": [2, 4, 6, 8, 10]})
result = causers.linear_regression(df, "x", "y")
print(f"y = {result.slope:.2f}x + {result.intercept:.2f}")

Features

  • Linear regression with HC3 robust standard errors

  • Logistic regression with Newton-Raphson MLE

  • Cluster-robust standard errors (analytical and bootstrap)

  • Fixed effects (within-transformation for OLS, Mundlak for logistic)

  • Synthetic Difference-in-Differences (SDID)

  • Synthetic Control (SC) with 4 method variants

  • Double Machine Learning (DML) with cross-fitting

  • Two-Stage Least Squares (IV/2SLS) with weak instrument diagnostics

  • Covariate balance diagnostics (SMD, variance ratios, weighted ESS)

Indices and tables