// the find
mrdbourke/zero-to-mastery-ml
All course materials for the Zero to Mastery Machine Learning and Data Science course.
This is the full notebook/data/slide set for Daniel Bourke's Zero to Mastery Machine Learning Udemy course, covering numpy, pandas, matplotlib, scikit-learn and a bit of TensorFlow/Keras through three milestone projects (heart disease classification, bulldozer price regression, dog breed image classification). It's for someone who wants a guided, linear path into practical ML with real messy datasets, not a library or anything you'd import into a project.
The three milestone projects use genuinely messy, real-world data (missing values, mixed types, imbalanced categories) instead of toy datasets like iris or MNIST, so the preprocessing steps are worth something. Datasets ship in the repo itself, so there's no separate download/Kaggle-auth step to get a notebook running. Binder and Colab badges mean you can run any notebook with zero local setup. The content is also mirrored into an mkdocs site, so you can read through without launching Jupyter at all.
There's no code to reuse here — it's entirely instructional notebooks, so the star count reflects course popularity, not a tool anyone adopts. The repo has accumulated duplicate and '-OLD' suffixed notebook variants (video vs non-video, v1 vs v2) that make it unclear which file is current without reading the README's update log. Only CI present is a docs-build workflow — no notebook execution tests, so a broken cell could sit unnoticed for a long time. The deep learning section is still just TensorFlow/Keras basics via transfer learning; there's no coverage of anything more current, and per the repo's own discussion thread the 2025 refresh was still in progress as of the last update.