BayesianBandits

Lightweight microframework for Bayesian bandits (Thompson Sampling) with support for contextual/restless/delayed rewards.

Adaptive Experimentation & Bandits

Docs GitHub

ContextualBandits

Implements a wide range of contextual bandit algorithms (linear, tree-based, neural) and off-policy evaluation methods.

Adaptive Experimentation & Bandits

Docs GitHub

MABWiser

Production-ready, scikit-learn style library for contextual & stochastic bandits with parallelism and simulation tools.

Adaptive Experimentation & Bandits

Docs GitHub

Open Bandit Pipeline (OBP)

Framework for **offline evaluation (OPE)** of bandit policies using logged data. Implements IPS, DR, DM estimators.

Adaptive Experimentation & Bandits

Docs GitHub

PyXAB

Library for advanced bandit problems: X-armed bandits (continuous/structured action spaces) and online optimization.

Adaptive Experimentation & Bandits

Docs GitHub

SMPyBandits

Comprehensive research framework for single/multi-player MAB algorithms (stochastic, adversarial, contextual).

Adaptive Experimentation & Bandits

Docs GitHub

Bambi

High-level interface for building Bayesian GLMMs, built on top of PyMC. Uses formula syntax similar to R's `lme4`.

Bayesian Econometrics

Docs GitHub

LightweightMMM

Bayesian Marketing Mix Modeling (see Marketing Mix Models section).

Bayesian Econometrics

GitHub

NumPyro

Probabilistic programming library built on JAX for scalable Bayesian inference, often faster than PyMC.

Bayesian Econometrics

Docs GitHub

PyMC

Flexible probabilistic programming library for Bayesian modeling and inference using MCMC algorithms (NUTS).

Bayesian Econometrics

Docs GitHub

Ananke

Causal inference using graphical models (DAGs), including identification theory and effect estimation.

Causal Discovery & Graphical Models

Docs GitHub

Causal Discovery Toolbox (CDT)

Implements algorithms for causal discovery (recovering causal graph structure) from observational data.

Causal Discovery & Graphical Models

Docs GitHub

CausalNex

Uses Bayesian Networks for causal reasoning, combining ML with expert knowledge to model relationships.

Causal Discovery & Graphical Models

GitHub

LiNGAM

Specialized package for learning non-Gaussian linear causal models, implementing various versions of the LiNGAM algorithm including ICA-based methods.

Causal Discovery & Graphical Models

Docs GitHub

Tigramite

Specialized package for causal inference in time series data implementing PCMCI, PCMCIplus, LPCMCI algorithms with conditional independence tests.

Causal Discovery & Graphical Models

Docs GitHub

causal-learn

Comprehensive Python package serving as Python translation and extension of Java-based Tetrad toolkit for causal discovery algorithms.

Causal Discovery & Graphical Models

Docs GitHub

gCastle

Huawei Noah's Ark Lab end-to-end causal structure learning toolchain emphasizing gradient-based methods with GPU acceleration (NOTEARS, GOLEM).

Causal Discovery & Graphical Models

Docs GitHub

py-tetrad

Python interface to Tetrad Java library using JPype, providing direct access to Tetrad's causal discovery algorithms with efficient data translation.

Causal Discovery & Graphical Models

GitHub

CausalInference

Implements classical causal inference methods like propensity score matching, inverse probability weighting, stratification.

Causal Inference & Matching

Docs GitHub

CausalLib

IBM-developed package that provides a scikit-learn-inspired API for causal inference with meta-algorithms supporting arbitrary machine learning models.

Causal Inference & Matching

Docs GitHub

CausalML

Focuses on uplift modeling and heterogeneous treatment effect estimation using machine learning techniques.

Causal Inference & Matching

Docs GitHub

Implements Propensity Score Matching (PSM) and Coarsened Exact Matching (CEM) with ML flexibility for propensity score estimation.

CausalPlayground

Python library for causal research that addresses the scarcity of real-world datasets with known causal relations. Provides fine-grained control over structural causal models.

Causal Inference & Matching

Docs GitHub

CausalPy

Developed by PyMC Labs, focuses specifically on causal inference in quasi-experimental settings. Specializes in scenarios where randomization is impossible or expensive.

Causal Inference & Matching

Docs GitHub

DoWhy

End-to-end framework for causal inference based on causal graphs (DAGs) and potential outcomes. Covers identification, estimation, refutation.

Causal Inference & Matching

Docs GitHub

Fast k-nearest-neighbor matching for large datasets using Facebook's FAISS library.

scikit-uplift

Focuses on uplift modeling and estimating heterogeneous treatment effects using various ML-based methods.

Causal Inference & Matching

Docs GitHub

Scikit-learn

Foundational ML library with regression models (incl. regularized), model selection, cross-validation, evaluation metrics.

Core Libraries & Linear Models

Docs GitHub

Statsmodels

Comprehensive library for estimating statistical models (OLS, GLM, etc.), conducting tests, and data exploration. Core tool.

Core Libraries & Linear Models

Docs GitHub

FactorAnalyzer

Specialized library for Exploratory (EFA) and Confirmatory (CFA) Factor Analysis with rotation options for interpretability.

Dimensionality Reduction

Docs GitHub

openTSNE

Optimized, parallel implementation of t-distributed Stochastic Neighbor Embedding (t-SNE) for large datasets.

Dimensionality Reduction

Docs GitHub

umap-learn

Fast and scalable implementation of Uniform Manifold Approximation and Projection (UMAP) for non-linear reduction.

Dimensionality Reduction

Docs GitHub

Biogeme

Maximum likelihood estimation of parametric models, with strong support for complex discrete choice models.

Discrete Choice Models

Docs GitHub

PyBLP

Tools for estimating demand for differentiated products using the Berry-Levinsohn-Pakes (BLP) method.

Discrete Choice Models

Docs GitHub

PyLogit

Flexible implementation of conditional/multinomial logit models with utilities for data preparation.

Discrete Choice Models

GitHub

XLogit

Fast estimation of Multinomial Logit and Mixed Logit models, optimized for performance.

Discrete Choice Models

Docs GitHub

torch-choice

PyTorch framework for flexible estimation of complex discrete choice models, leveraging GPU acceleration.

Discrete Choice Models

Docs GitHub

DoubleML

Implements the double/debiased ML framework (Chernozhukov et al.) for estimating causal parameters (ATE, LATE, POM) with ML nuisances.

Double/Debiased Machine Learning (DML)

Docs GitHub

EconML

Microsoft toolkit for estimating heterogeneous treatment effects using DML, causal forests, meta-learners, and orthogonal ML methods.

Double/Debiased Machine Learning (DML)

Docs GitHub

Double‑post Lasso estimator for high‑dimensional treatment effects (Belloni‑Chernozhukov‑Hansen 2014).

Debiased‑Lasso detector of heterogeneous treatment effects in randomized experiments.

Lightweight package for setting up and estimating custom GMM models based on user-defined moment conditions.

Lifetimes

Analyze customer lifetime value (CLV) using probabilistic models (BG/NBD, Pareto/NBD) to predict purchases.

Marketing Mix Models (MMM) & Business Analytics

Docs GitHub

MaMiMo

Lightweight Python library focused specifically on Marketing Mix Modeling implementation.

Marketing Mix Models (MMM) & Business Analytics

GitHub

PyMC Marketing

Collection of Bayesian marketing models built with PyMC, including MMM, CLV, and attribution.

Marketing Mix Models (MMM) & Business Analytics

Docs GitHub

mmm_stan

Python/STAN implementation of Bayesian Marketing Mix Models.

Marketing Mix Models (MMM) & Business Analytics

GitHub

Gensim

Library focused on topic modeling (LDA, LSI) and document similarity analysis.

Natural Language Processing for Economics

Docs GitHub

Transformers

Access to thousands of pre-trained models for NLP tasks like text classification, summarization, embeddings, etc.

Natural Language Processing for Economics

Docs GitHub

spaCy

Industrial-strength NLP library for efficient text processing pipelines (NER, POS tagging, etc.).

Natural Language Processing for Economics

Docs GitHub

JAX

High-performance numerical computing with autograd and XLA compilation on CPU/GPU/TPU.

Numerical Optimization & Computational Tools

Docs GitHub

PyTorch

Popular deep learning framework with flexible automatic differentiation.

Numerical Optimization & Computational Tools

Docs GitHub

Solves linear models with high-dimensional fixed effects, supporting robust variance calculation and IV.

Linearmodels

Estimation of fixed, random, pooled OLS models for panel data. Also Fama-MacBeth and between/first-difference estimators.

Panel Data & Fixed Effects

Docs GitHub

Fast estimation of linear models with multiple high-dimensional fixed effects (like R's `fixest`). Supports OLS, IV, Poisson, robust/cluster SEs.

Out-of-core regression (OLS/IV) for very large datasets using DuckDB aggregation. Handles data that doesn't fit in memory.

pydynpd

Estimation of dynamic panel data models using Arellano-Bond (Difference GMM) and Blundell-Bond (System GMM). Includes Windmeijer correction & tests.

Panel Data & Fixed Effects

Docs GitHub

ADOpy

Bayesian Adaptive Design Optimization (ADO) for tuning experiments in real-time, with models for psychometric tasks.

Power Simulation & Design of Experiments

Docs GitHub

Adaptive

Parallel active learning library for adaptive function sampling/evaluation, with live plotting for monitoring.

Power Simulation & Design of Experiments

Docs GitHub

DoEgen

Automates generation and optimization of designs, especially for mixed factor-level experiments; computes efficiency metrics.

Power Simulation & Design of Experiments

GitHub

pyDOE2

Implements classical Design of Experiments: factorial (full/fractional), response surface (Box-Behnken, CCD), Latin Hypercube.

Power Simulation & Design of Experiments

Docs GitHub

CausalImpact

Python port of Google's R package for estimating causal effects of interventions on time series using Bayesian structural time-series models.

Program Evaluation Methods (DiD, SC, RDD)

Docs GitHub

Differences

Implements modern difference-in-differences methods for staggered adoption designs (e.g., Callaway & Sant'Anna).

Program Evaluation Methods (DiD, SC, RDD)

Docs GitHub

SyntheticControlMethods

Implementation of synthetic control methods for comparative case studies when panel data is available.

Program Evaluation Methods (DiD, SC, RDD)

GitHub

csdid

Python adaptation of the R `did` package. Implements multi-period DiD with staggered treatment timing (Callaway & Sant’Anna).

Program Evaluation Methods (DiD, SC, RDD)

GitHub

mlsynth

Implements advanced synthetic control methods: forward DiD, cluster SC, factor models, and proximal SC. Designed for single-treated-unit settings.

Program Evaluation Methods (DiD, SC, RDD)

Docs GitHub

Changes‑in‑Changes (CiC) estimator for distributional treatment effects (Athey & Imbens 2006).

Lee (2009) sample‑selection bounds for treatment effects; trims treated distribution to match selection rates.

rdd

Toolkit for sharp RDD analysis, including bandwidth calculation and estimation, integrating with pandas.

Program Evaluation Methods (DiD, SC, RDD)

GitHub

rdrobust

Comprehensive tools for Regression Discontinuity Designs (RDD), including optimal bandwidth selection, estimation, inference.

Program Evaluation Methods (DiD, SC, RDD)

Docs GitHub

Fast quantile regression solver using interior point methods, supporting robust and clustered standard errors.

Recentered Influence‑Function (RIF) regression for unconditional quantile & distributional effects (Firpo et al., 2008).

quantile-forest

Scikit-learn compatible implementation of Quantile Regression Forests for non-parametric estimation.

Quantile Regression & Distributional Methods

Docs GitHub

(PySAL Core)

The broader PySAL ecosystem contains many tools for spatial data handling, weights, visualization, and analysis.

Spatial Econometrics

Docs GitHub

PySAL (spreg)

The spatial regression `spreg` module of PySAL. Implements spatial lag, error, IV models, and diagnostics.

Spatial Econometrics

Docs GitHub

Curated list of quantitative finance libraries and resources (many statistical/TS tools overlap with econometrics).

Teaches software design principles for ML—modularity, abstraction, and reproducibility—going beyond ad hoc Jupyter workflows. Focus on maintainable, production-quality ML code.

Modern introduction to causal inference methods (DiD, IV, RDD, Synth, ML-based) with Python code examples.

Practical guide by A. Turrell on using Python for modern econometric research, data analysis, and workflows.

Intermediate 5-course series by Andrew Ng covering deep neural networks, CNNs, RNNs, transformers, and real-world DL applications using TensorFlow.

Beginner-friendly 3-course series by Andrew Ng covering core ML methods (regression, classification, clustering, trees, NN) with hands-on projects.

Comprehensive intro notes by Kevin Sheppard covering Python basics, core libraries, and econometrics applications.

High-quality lecture series on quantitative economic modeling, computational tools, and economics using Python/Julia.

SciPy Bootstrap

(`scipy.stats.bootstrap`) Computes bootstrap confidence intervals for various statistics using percentile, BCa methods.

Standard Errors, Bootstrapping & Reporting

Docs GitHub

Stargazer

Python port of R's stargazer for creating publication-quality regression tables (HTML, LaTeX) from `statsmodels` & `linearmodels` results.

Standard Errors, Bootstrapping & Reporting

GitHub

Teaches essential developer tools often skipped in formal education—command line, Git, Vim, scripting, debugging, etc.

wildboottest

Fast implementation of various wild cluster bootstrap algorithms (WCR, WCU) for robust inference, especially with few clusters.

Standard Errors, Bootstrapping & Reporting

Docs GitHub

FilterPy

Focuses on Kalman filters (standard, EKF, UKF) and smoothers with a clear, pedagogical implementation style.

State Space & Volatility Models

Docs GitHub

Metran

Specialized package for estimating Dynamic Factor Models (DFM) using state-space methods and Kalman filtering.

State Space & Volatility Models

GitHub

PyKalman

Implements Kalman filter, smoother, and EM algorithm for parameter estimation, including support for missing values and UKF.

State Space & Volatility Models

Docs GitHub

PyMC Statespace

(See Bayesian) Bayesian state-space modeling using PyMC, integrating Kalman filtering within MCMC for parameter estimation.

State Space & Volatility Models

Docs GitHub

stochvol

Efficient Bayesian estimation of stochastic volatility (SV) models using MCMC.

State Space & Volatility Models

Docs GitHub

Pingouin

User-friendly interface for common statistical tests (ANOVA, ANCOVA, t-tests, correlations, chi², reliability) built on pandas & scipy.

Statistical Inference & Hypothesis Testing

Docs GitHub

PyWhy-Stats

Part of the PyWhy ecosystem providing statistical methods specifically for causal applications, including various independence tests and power-divergence methods.

Statistical Inference & Hypothesis Testing

Docs GitHub

Scipy.stats

Foundational module within SciPy for a wide range of statistical functions, distributions, and hypothesis tests (t-tests, ANOVA, chi², KS, etc.).

Statistical Inference & Hypothesis Testing

Docs GitHub

hypothetical

Library focused on hypothesis testing: ANOVA/MANOVA, t-tests, chi-square, Fisher's exact, nonparametric tests (Mann-Whitney, Kruskal-Wallis, etc.).

Statistical Inference & Hypothesis Testing

GitHub

lifelines

Comprehensive library for survival analysis: Kaplan-Meier, Nelson-Aalen, Cox regression, AFT models, handling censored data.

Statistical Inference & Hypothesis Testing

Docs GitHub

Dolo

Framework for describing and solving economic models (DSGE, OLG, etc.) using a declarative YAML-based format.

Structural Econometrics & Estimation

Docs GitHub

HARK

Toolkit for solving, simulating, and estimating models with heterogeneous agents (e.g., consumption-saving).

Structural Econometrics & Estimation

Docs GitHub

QuantEcon.py

Core library for quantitative economics: dynamic programming, Markov chains, game theory, numerical methods.

Structural Econometrics & Estimation

Docs GitHub

respy

Simulation and estimation of finite-horizon dynamic discrete choice (DDC) models (e.g., labor/education choice).

Structural Econometrics & Estimation

Docs GitHub

SDV (Synthetic Data Vault)

Comprehensive library for generating synthetic tabular, relational, and time series data using various models.

Synthetic Data Generation

Docs GitHub

Synthpop

Port of the R package for generating synthetic populations based on sample survey data.

Synthetic Data Generation

GitHub

ARCH

Specialized library for modeling and forecasting conditional volatility using ARCH, GARCH, EGARCH, and related models.

Time Series Econometrics

Docs GitHub

Kats

Broad toolkit for time series analysis, including multivariate analysis, detection (outliers, change points, trends), feature extraction.

Time Series Econometrics

Docs GitHub

LocalProjections

Community implementations of Jordà (2005) Local Projections for estimating impulse responses without VAR assumptions.

Time Series Econometrics

GitHub

MLForecast

Scalable time series forecasting using machine learning models (e.g., LightGBM, XGBoost) as regressors.

Time Series Forecasting

Docs GitHub

NeuralForecast

Deep learning models (N-BEATS, N-HiTS, Transformers, RNNs) for time series forecasting, built on PyTorch Lightning.

Time Series Forecasting

Docs GitHub

Prophet

Forecasting procedure for time series with strong seasonality and trend components, developed by Facebook.

Time Series Forecasting

Docs GitHub

StatsForecast

Fast, scalable implementations of popular statistical forecasting models (ETS, ARIMA, Theta, etc.) optimized for performance.

Time Series Forecasting

Docs GitHub

pmdarima

ARIMA modeling with automatic parameter selection (auto-ARIMA), similar to R's `forecast::auto.arima`.

Time Series Forecasting

Docs GitHub

sktime

Unified framework for various time series tasks, including forecasting with classical, ML, and deep learning models.

Time Series Forecasting

Docs GitHub

CatBoost

Gradient boosting library excelling with categorical features (minimal preprocessing needed). Robust against overfitting.

Tree & Ensemble Methods for Prediction

Docs GitHub

LightGBM

Fast, distributed gradient boosting (also supports RF). Known for speed, low memory usage, and handling large datasets.

Tree & Ensemble Methods for Prediction

Docs GitHub

NGBoost

Extends gradient boosting to probabilistic prediction, providing uncertainty estimates alongside point predictions. Built on scikit-learn.

Tree & Ensemble Methods for Prediction

Docs GitHub

Scikit-learn Ens.

(`RandomForestClassifier`/`Regressor`) Widely-used, versatile implementation of Random Forests. Easy API and parallel processing support.

Tree & Ensemble Methods for Prediction

Docs GitHub

XGBoost

High-performance, optimized gradient boosting library (also supports RF). Known for speed, efficiency, and winning competitions.

Tree & Ensemble Methods for Prediction

Docs GitHub

cuML (RAPIDS)

GPU-accelerated implementation of Random Forests for significant speedups on large datasets. Scikit-learn compatible API.

Tree & Ensemble Methods for Prediction

Docs GitHub