An SBB Forschungsfonds project that tackles a concrete problem. SBB’s Synthetic Population (SynPop) — 9 million persons, 4 million households and 700,000 businesses, geo-referenced to exact coordinates — is one of the most detailed pictures of Switzerland ever assembled for transport planning. Yet it can barely be shared, because it is built on confidential BFS register data (STATPOP, STATENT).
We develop household-aware anonymization for the register-derived variables — including geographic protection of household coordinates — using targeted record swapping and geographic perturbation. We pair this with a formal disclosure-risk assessment that goes beyond k-anonymity, to attribution and predictive-disclosure risk and geographic re-identification, so that BFS and SBB gain an evidence-based, differentiated release policy instead of today’s all-or-nothing.
The aim is individual-level population data that can finally be used safely in agent-based transport models such as SIMBA MOBi and the National Passenger Transport Model (NPVM). The project builds on a 2025 proof-of-concept for the canton of Ticino and is led by Prof. Dr. Matthias Templ (FHNW), with Oscar Thees and Roman Müller on the project team, in collaboration with SBB and the Swiss Federal Statistical Office (BFS).
