ESTP – Differential Privacy for Official Statistics

This repository contains Jupyter notebooks for the ESTP course on Differential Privacy, with a focus on practical tools and hands-on exercises.
The material introduces how to apply differential privacy using open-source libraries.

---

⚙️ Setup Instructions

1. Clone this repository:

   git clone https://github.com/dscc-admin-ch/ESTP_DP_Course.git
   cd ESTP_DP_Course
   
   

2. Install global dependencies

   pip install -r requirements.txt
   
   

3. Launch Jupyter: Open a notebook and follow the instructions!

---

📚 Repo Structure

The repository is organized around three main libraries for differential privacy:

1. SmartNoise SQL
   • Apply differential privacy to SQL queries.
2. Diffprivlib
   • IBM’s library for differentially private machine learning and statistics.
3. OpenDP
   • The OpenDP project’s core library for building custom DP transformations and working with polars tables.

For each library, you will find:

• Exercise notebook – practical problems to solve.
• Correction notebook – worked-out solutions.

📂 Repository Layout

ESTP_DP_Course/
│
├── smartnoise-sql/
│   ├── Smartnoise-SQL-Exercises.ipynb
│   └── Smartnoise-SQL-Corrections.ipynb
│
├── diffprivlib/
│   ├── DiffPrivLib-Exercises.ipynb
│   └── DiffPrivLib-Corrections.ipynb
│
├── opendp_polars/
│   ├── OpenDP_Polars_Exercises.ipynb
│   └── OpenDP_Polars_Corrections.ipynb
│
└── README.md

---

📝 Notes

• Exercises are designed for practice during the course.
• Corrections provide detailed solutions — use them for self-study or after attempting the exercises.
• Make sure you install the correct version of each library.

---

📖 References

• OpenDP Project
• SmartNoise Documentation
• IBM Diffprivlib
• Differential Privacy Overview (Wikipedia)

---

Happy learning! 🚀