When building tree-based models, such as CART, random forests, or XGBoost, we might be interested in knowing a little more about how the model works. For example, if a CART and XGBoost model had roughly the same performance, we might want to characterize how complex each is so that we are more informed about which to prefer. Knowing how many predictors were used, how many terminal nodes are in the tree, and similar characteristics can help understand the model. We might desire to visualize the tree (or a tree in the ensemble), and so on.
This is a companion discussion topic for the original entry at https://opensource.posit.co/blog/2026-07-28_lorax/