UKDRI GitHub Training — GitHub Actions for MLOps

Session date: 16 April 2026

This is the reference repository for the UKDRI GitHub training session on setting up GitHub Actions for MLOps. It will be used as a live demo during the session — a new workflow .yml file will be created from scratch and wired up to this pipeline.

---

The ML Problem

A logistic regression model trained to predict status — whether a subject has Parkinson's disease (1) or is healthy (0) — from 22 voice acoustic features (jitter, shimmer, harmonics-to-noise ratio, and related measures).

Dataset: UCI Parkinson's Disease Data Set (195 subjects, open-access).
Downloaded at runtime by data/download_data.py, which falls back to OpenML if UCI is unavailable.

---

Project Structure

.github/workflows/   — workflow definitions
data/                — dataset download script
src/                 — train.py, evaluate.py, predict.py
tests/               — data integrity and model quality tests
params.yaml          — model hyperparameters
requirements.txt

---