MIDAS2Multiple Imputation with Denoising Autoencoders
The next generation of MIDAS: a deep learning method for accurate and efficient multiple imputation of missing data, available for both Python and R.
My open-source projects translate methodological research into tools for applied scientists.
Packages and resources
citestConditional Independence of Missingness Test
Implementation of the PAIR-CI conditional independence test for incomplete data, available for both Python and R.
cjrankRanked-Choice Conjoint Experiments
An R package for estimating effects from conjoint experiments in which respondents rank, rather than choose between, profiles.
cjbartHeterogeneous Effects Analysis of Conjoint Experiments
An R package for analysing conjoint experiments using Bayesian Additive Regression Trees (BART), focused on inspecting heterogeneous treatment effects.
SyGNetSynthetic Data via Generative Adversarial Networks
A Python package for generating synthetic data for the social sciences using deep learning methods.
10 Fundamental Theorems for Econometrics
Online textbook covering ten foundational theorems in econometrics.