Help a clinic plan its day
Study appointment attendance with a careful, responsible approach.
What can a small appointment dataset tell us about attendance?
Use a synthetic appointment dataset to study non-attendance. The goal is to understand model evaluation and potential service differences. A real clinic would need approved data governance, stakeholder review, and prospective evaluation before using a model.
Build it, step by step
- 01
Inspect the data, class balance, and appointment lead times. All records are synthetic and contain no patient data. Treat appointment_id as an identifier.
- 02
Hold out a stratified final test set. Use lead_days and previous_missed as features; the fictional service_group column is reserved for auditing.
- 03
Train a logistic regression pipeline using training data only. Compare it with a majority-class baseline through cross-validation.
- 04
Use the same final threshold for both fictional groups. On the test set, compare false negative rates and report the sample count in each group. Explain why small samples make differences uncertain.
- 05
Write a model card covering intended use, data limits, errors, group results, and why this model must not determine access to care. Propose a supportive intervention to study, such as optional appointment reminders.
Your project checklist
- An untouched final test set and a reproducible training pipeline.
- Overall precision, recall, and a confusion matrix.
- Group error rates shown alongside denominators, with uncertainty discussed.
- A model card stating that this is a teaching exercise and cannot guide patient care or access.
Free, open and run by a nonprofit.
This page is part of Python to AI, a free course of 20 modules and 60 lessons in the engineering.fit( ) Learning Lab. The Learning Lab is built and run by engineering.fit( ), a nonprofit founded by Abdullateef Opeyemi Bakare that helps African engineering students gain research experience. You do not need an account or pay a fee to learn.