// the find
campusx-official/100-days-of-machine-learning
A 100-day structured ML curriculum delivered as Jupyter notebooks, covering the full pipeline from pandas basics through ensemble methods. Aimed squarely at beginners who learn by running code, not reading textbooks. The companion YouTube series from CampusX is what gives the notebooks their context.
Each day is self-contained with its own folder, dataset, and readme — you can drop into any topic without running everything before it. The gradient descent section (day51-52) is unusually thorough: multiple notebooks animating batch vs SGD vs mini-batch with actual matplotlib animations, not just static plots. sklearn pipelines get a full treatment across days 28-30 including serialized models, which most intro courses skip entirely. Data files are committed alongside the notebooks so nothing breaks on first run.
The folder naming is inconsistent — some use spaces, some use hyphens, day23 is missing, day41 is missing, and there are orphaned files like adaboost_demo.ipynb floating at the root next to day66 which has the same content. Last commit was October 2024 and the curriculum stops well short of 100 days of content (no deep learning, no NLP beyond tokenization basics). No requirements.txt or environment file anywhere — version mismatches between sklearn, pandas, and matplotlib will silently break cells.