Software

My research group develops open-source software that translates methodological advances in causal and statistical inference into tools that researchers can use in practice.

R package

flexCausal

Causal effectsHidden/Unmeasured/Latent variablesTMLE/One-step estimatorDouble-Debiased ML

Flexible estimation of non-parametrically identifiable average causal effects in graphical models with unmeasured variables, including one-step corrected plug-in and targeted minimum loss-based estimators.

GitHub CRAN

R package

flexPaths

Causal mediationCausal path-specific effectsDouble-Debiased MLLongitudinal studies

Flexible estimation of path-specific causal effects in mediation analyses with multiple ordered mediators and multiple treatments. Supports influence-function based estimation and flexible pathway decompositions.

GitHub

R package

flexMissing

Missing dataGraphical modelsMissing-not-at-randomIdentification and estimation

Weighting-based identification and estimation of target parameters in graphical models with missing data, supporting MCAR, MAR, and MNAR settings. Flexible paraemter estimators can also leverage machine learning methods.

GitHub

Python package

Ananke

Graphical modelsIdentificationParametric estimationLinear SEMs

Tools for causal inference with graphical models, including constructing and visualizing causal graphs, identifying causal effects, and performing (semi)parametric estimation across a variety of causal structures.

GitLab Documentation Package Paper

R package

coxphMIC

Survival analysisSparse estimation

Sparse estimation of Cox proportional hazards models via approximated information criteria.

R Repository Package Paper