Program and Agenda
Detailed schedule with hands-on blocks, breaks, and lab prerequisites.
Full-day schedule (Wednesday, April 22, 2026)
| Time (CET) | Session | Covers | Format |
|---|---|---|---|
| 09:00-09:20 | Welcome and framing | Scope, expectations, learning goals | Plenary |
| 09:20-10:10 | Definitions and risk basics | anonymization vs pseudonymization, direct/indirect identifiers, linkage risk | Lecture + discussion |
| 10:10-10:30 | Break | Coffee and setup checks | Break |
| 10:30-11:30 | Threat modeling in practice | attacker models, auxiliary data, contextual risk assumptions | Guided exercise |
| 11:30-12:30 | Lab 1: Tabular anonymization | k-anonymity, l-diversity, t-closeness on sample healthcare-style data | Hands-on lab |
| 12:30-13:30 | Lunch | Networking and Q&A | Break |
| 13:30-14:20 | Utility and release quality | utility metrics, bias checks, decision logs | Lecture + demo |
| 14:20-15:10 | Differential privacy essentials | epsilon intuition, mechanism selection, deployment constraints | Lecture + demo |
| 15:10-15:25 | Break | Reset before second lab | Break |
| 15:25-16:20 | Lab 2: Privacy-utility tradeoff | compare anonymization variants and document release decision | Hands-on lab |
| 16:20-16:45 | Synthetic data and residual risk | when synthetic helps, where risk remains | Discussion |
| 16:45-17:00 | Wrap-up and next steps | implementation checklist, resources, post-workshop path | Plenary |
Expected end time: 17:00 CET.
Session-specific prerequisites
- Lab 1: Python 3.11+, JupyterLab, and the anonymization notebook bundle from the Materials setup section.
- Lab 2: Same setup as Lab 1, plus local Docker access or cloud notebook fallback.
- All sessions: participants should know basic SQL or pandas-style data operations.
Hands-on labeling
- Sessions marked Hands-on lab require your local setup to be completed in advance.
- Instructors and support staff provide troubleshooting during breaks and lab start.