Zurich's municipal digital infrastructure is facing a reckoning. Across several city departments and public institutions, administrators have been quietly grappling with a sprawling problem: thousands of duplicate images sitting in overlapping archives, consuming server capacity, distorting search results, and complicating efforts to modernise the city's data systems ahead of a planned digital services overhaul scheduled for 2027.
The issue is not unique to Zurich — cities from Amsterdam to Vienna have run into similar headaches as legacy systems built in the 2000s collide with modern cloud-based platforms. But the stakes here are sharper than most. The city is already under pressure to demonstrate fiscal discipline following the turbulence in Swiss banking over the past three years, and throwing money at a botched data migration would land badly at the Stadtrat, where budget oversight has grown notably more vigilant.
Where the Problem Lives — and Why It Has Taken This Long
The duplication is concentrated in two main areas: the Stadtarchiv, Zurich's official municipal archive on Neumarkt, and the digital collections managed through the Zentralbibliothek on Zähringerplatz. Both institutions digitised large portions of their photographic holdings independently, often around the same periods, without a shared deduplication protocol. The result is a patchwork of cataloguing standards, inconsistent metadata, and in some cases the same image stored under different file names across both systems.
A working group drawn from Stadt Zürich's Informatik-Dienste and representatives from both institutions has been meeting since February 2026 to map the scale of the problem. The group has not yet published a final count of affected files, but internal planning documents reviewed by The Daily Zurich suggest the duplication rate in certain photographic collections runs well into the tens of thousands of individual image files. That volume matters because any automated deduplication tool needs to be trained carefully enough to distinguish genuine duplicates from near-identical images taken seconds apart — a common occurrence in historical event photography.
ETH Zurich's Data Science group has been consulted informally on the technical approach, specifically around perceptual hashing algorithms that can identify visually similar images without relying on metadata alone. Whether a formal contract follows depends on procurement decisions expected before the end of the third quarter this year.
The Decisions That Will Define the Outcome
Three choices are coming to a head before autumn. First, the city must decide whether to build or buy: develop a bespoke deduplication pipeline using internal IT capacity, or license an existing platform. A commercial licence for an enterprise image-management solution capable of handling archives of this scale typically runs between CHF 80,000 and CHF 200,000 annually depending on volume tiers — a figure that would need sign-off through the standard Stadtrat expenditure process.
Second, whoever handles the technical work will need a clear policy on what happens to confirmed duplicates. Deleting files outright from public archives carries legal and cultural heritage risks. The more cautious approach — flagging duplicates and moving them to cold storage — costs less to execute politically, but adds ongoing infrastructure expense. The Stadtarchiv's mandate under cantonal law to preserve records indefinitely complicates any aggressive deletion strategy.
Third is the question of governance going forward. Without a unified metadata standard and a shared ingest protocol between the Stadtarchiv and the Zentralbibliothek, the same problem will rebuild itself within five years. Establishing a joint digital stewardship committee, with enforced technical standards for any new image acquisitions, is the structural fix — but it requires both institutions to cede some autonomy over their own cataloguing practices, which historically has been a sticking point.
A preliminary report from the working group is expected in September 2026, ahead of budget deliberations for the 2027 fiscal year. How the Stadtrat responds — and how much it is willing to commit before the wider digital infrastructure overhaul begins — will set the terms for everything that follows. The summer recess buys a little time, but not much.