finds.dev← search

// the find

susanli2016/Machine-Learning-with-Python

★ 4,592 · Jupyter Notebook · updated Jun 2025

Python code for common Machine Learning Algorithms

A flat dump of ~100 Jupyter notebooks covering classical ML algorithms and applied data science projects, mostly from Towards Data Science articles by Susan Li. It's a personal portfolio masquerading as a learning resource — the notebooks are real work on real datasets, not toy examples.

- Covers practical applied ML across domains nobody usually touches together: hospitality, CLV modeling, election forecasting, survival analysis — not just the standard Iris/Titanic examples.

- Bayesian methods get real treatment: PyMC3, PyStan, hierarchical regression with post-stratification — not just a scikit-learn LogisticRegression wrapper.

- Recommender systems section is unusually thorough: BPR with implicit feedback, collaborative filtering, hybrid approaches with TensorFlow — most intro repos skip implicit feedback entirely.

- The notebooks appear to have been actually run and saved with outputs, so you can follow the reasoning without executing anything.

- Zero organization: 100+ notebooks in a flat root directory with no index, no grouping by topic, no progression from beginner to advanced. Finding anything requires grep.

- Dependencies are undeclared — no requirements.txt, no environment.yml, no pinned versions. Half these notebooks will fail on import because the library APIs have changed since they were written.

- Last meaningful update was years ago for most notebooks; the 2025 push date doesn't reflect the content age. PyMC3 is now PyMC, TensorFlow 1.x patterns appear, some notebooks reference deprecated Pandas APIs.

- It's one person's published articles saved as notebooks — the 'teaching' is incidental. No exercises, no explanations written for someone learning, no attempt to build on prior notebooks.

View on GitHub →

// want more like this?

We dig through GitHub every week and send a few repos picked for what you actually care about — each with an honest take like this one.

Get finds in your inbox → Search again →