// the find
justmarkham/scikit-learn-videos
Jupyter notebooks from the scikit-learn video series
Ten Jupyter notebooks accompanying a YouTube video series on scikit-learn basics — classification, regression, cross-validation, grid search, pipelines. Aimed squarely at beginners who want a structured, video-paired introduction to applied machine learning in Python.
The notebook-to-video pairing is genuinely useful: you can pause, run code, and follow along without jumping between a repo and slides. Coverage of cross-validation and grid search together in consecutive notebooks builds the mental model correctly rather than treating them as separate topics. The Pipeline + ColumnTransformer finale in notebook 10 is the right place to end — it's where beginners usually get lost and the example is concrete.
The notebooks are pinned to scikit-learn 0.23.2 (released 2020); the library is now at 1.5.x and the API has changed in places, so some code will raise deprecation warnings or break outright. There are only two tiny CSV datasets bundled, which means you can't experiment much without going off-script. The series stops at the 'how does scikit-learn work' level and never touches anything from the last five years of practice — no feature importance, no SHAP, no handling of imbalanced classes. If you already know the scikit-learn fit/predict pattern, there's nothing here for you.