This is a learning project to understand Generalised Linear Models (GLM) using Python.
→ Watch it directly on YouTube
Use Python 3.12, create a virtual environment, and install requirements.txt.
From the repository root, run python -m unittest discover -s tests to exercise
fitting, four diagnostic plots and model export on seeded synthetic test data.
For a headless environment, set MPLBACKEND=Agg.
python src/glm_pipeline.py reads data/sample.csv and writes
models/glm_model.pickle. The bundled three-row CSV only illustrates the schema:
it is insufficient for the four model parameters and is rejected explicitly.
Provide a larger dataset with the same columns before running the pipeline.
Having more rows than parameters is only a basic guard; assess separation,
rank, convergence and model assumptions before interpreting results.
The notebook is an unimplemented placeholder. Dependencies are unpinned, so installation is not yet reproducible across time. These examples demonstrate software behaviour, not validated real-world credit-risk predictions.
Pierre-Henry Soria — An Extremely Passionate Data Scientist who loves automating content creation efficiently with meaningful data! Also a true roquefort 🧀, ristretto ☕️, and dark chocolate lover! 😋
Enthusiast of YouTube, AI, learning, and researching! Find me at pH7.me
Do you enjoy this project? Buy me a coffee (spoiler: I love almond extra-hot flat white coffees).
Distributed under the MIT License 🎉 Happy hacking! 🤠