Causal Machine Learning for Discovering Heterogeneous Treatment Effects

This repository contains the lecture slides and code tutorial from the ACIC 2026 short course on ``Causal Machine Learning for Discovering Heterogeneous Treatment Effects''.

• slides/: contains short course lecture slides
• R/: contains R helper functions for coding demonstration
• notebooks/: contains quarto/R markdown notebooks for coding demonstration
  • demo.qmd: quarto notebook with source code for coding walkthrough
  • demo.html: rendered html output from demo.qmd
  • demo.Rmd: R markdown notebook with source code for coding walkthrough; this is provided for users who prefer R markdown over quarto

R Packages/Dependencies

To run the demo.qmd (or demo.Rmd) notebook, you will need to have the following tools installed on your computer:

• quarto: a scientific and technical publishing system built on Pandoc; you can install quarto here
• R
  • The code was initially developed using version 4.4.1, but other versions of R will likely work

All R packages can be installed from the renv.lock file. To do so,

1. Open the acic2026-hte-short-course.Rproj file in RStudio

2. Check that the renv R package has been installed on your computer. If not, you can install it via: install.packages('renv')

3. After installing the renv R package, run the following line of code in your R console:

   renv::restore()

   This will install all R packages specified in the renv.lock file, which should be all the R packages needed to run the code in this repository.

Alternatively, you can install the necessary R packages manually. The R packages used in this repository include:

• yaml
• rmarkdown
• here
• dplyr
• ggplot2
• grf
• xnie/rlearner
• tibble
• tidyr
• tidyselect
• GGally
• causalDT
• CRE
• knitr
• lmtest
• sandwich
• glmnet
• xgboost (version 1.7.8.1)