Hospital readmission risk for diabetes patients

The brief

Readmissions cost hospitals money and harm patients. Using a decade of hospital encounters, predict 30-day readmission for diabetes patients, find the drivers (medications, prior visits, discharge type), and design a follow-up programme for the highest-risk patients. Good looks like: a calibrated risk model, drivers explained, and a costed follow-up plan.

What you will build

1. Clean the encounter data and define the 30-day label
2. Engineer features from diagnoses, medications and history
3. Train and calibrate models
4. Explain the drivers
5. Design and cost a follow-up programme
6. Write the report for a hospital board

Free, open data to build it with

• Diabetes 130-US Hospitals (1999–2008) ↗ UCI Machine Learning Repository

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.