Batch quality failure predictor

The brief

Failed batches mean scrapped product and investigations. Using real tablet manufacturing data, predict which batches are likely to fail quality checks from process settings and raw-material properties, and use the model to point investigators at likely causes. Good looks like: careful validation by batch, metrics for a rare outcome, and explanations a process engineer agrees with.

What you will build

1. Join process, material and quality data by batch
2. Define the failure outcome
3. Train gradient boosting with imbalance handling
4. Evaluate with recall and precision
5. Explain likely causes with SHAP

Free, open data to build it with

• Pharmaceutical manufacturing process and quality data ↗ Scientific Data (Nature)

Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.