gmonce/scikit-learn-book ? reverse-engineered prompt

Build me a small learning repo for the scikit learn book, with Jupyter notebooks that walk through the main machine learning topics in a friendly way. I want chapters for a gentle introduction, supervised learning with examples like Titanic, image recognition, regression, and text classification, plus unsupervised learning with clustering and PCA, and a couple of notebooks on feature engineering and model selection.

Please include any sample data and images needed so the notebooks run smoothly, and make the explanations beginner friendly, like something you could follow even if you are new to machine learning. If any details are unclear, look up current scikit learn docs online and use sensible defaults. Keep it organized so each notebook feels like a self contained lesson, and make sure the code is clean enough for someone to read and rerun from top to bottom.