dtreeviz

Why we included this project

Explaining what a trained tree actually decided usually means squinting at text rules or crude default renderings. This library draws real decision tree diagrams with the feature value distributions at each split, marks the path a given prediction takes, and adds leaf details plus class boundary plots for classifiers. The same adapter-style API covers scikit-learn, XGBoost, LightGBM, and Spark MLlib, so the way you inspect a model stays consistent even when the framework changes. That makes it practical for showing a stakeholder why a specific prediction came out the way it did, and for catching nonsense before a model goes anywhere near production.