How Zurich's Digital Archives Ended Up Full of Duplicate Images — and Why Fixing It Took a Decade

Zurich's public institutions collectively hold millions of digital images — architectural surveys, urban planning renders, social services case files, transit documentation — and a substantial share of them are exact or near-exact copies of each other. That is not a rumour from inside a data centre. It is the documented outcome of more than a decade of siloed, uncoordinated digitisation across cantonal departments, municipal agencies, and federally funded research bodies.

The problem matters today because Zurich is mid-way through a digital infrastructure consolidation that the canton began formalising in 2022 under its eGovernment strategy framework. Cleaning up duplicate image data is not a housekeeping footnote. It directly affects storage costs, search reliability, and the legal traceability of documents used in planning decisions — including the contentious housing approvals that have defined Zurich politics for the better part of three years.

How the Duplication Built Up

The roots go back to roughly 2010, when individual departments began digitising paper archives independently. The Stadtarchiv Zürich on Neumarkt, which holds records going back centuries, ran its own scanning programme. So did the Amt für Städtebau, responsible for urban development imagery. ETH Zurich's library system, headquartered on Rämistrasse, operated separate ingestion pipelines for research photography and scientific illustration. None of these systems spoke to each other in any automated way.

When a document travelled between departments — a building permit application that started at the Amt für Baubewilligungen and was reviewed by multiple other agencies — its attached images were frequently re-uploaded rather than referenced. Each upload created a new file. Multiply that by years of inter-departmental correspondence and the duplication compounds quickly. By the time the canton commissioned an audit of its central data warehouse in 2023, internal assessments reportedly found duplication rates in some image categories running well above 30 percent, though the canton has not published a definitive public figure.

The problem was not unique to Zurich. Bern and Basel-Stadt faced comparable challenges as they digitised. But Zurich's scale — it is Switzerland's largest canton by population, with roughly 1.6 million residents — made the volume unusually large and the remediation unusually expensive.

What Changed, and When

The inflection point came in late 2024, when the canton's IT directorate, the Amt für Informatik, began piloting duplicate-detection software across a subset of planning and social services imagery. The pilot drew on perceptual hashing techniques — algorithms that generate a compact fingerprint for each image and flag matches even when files have been slightly resized or recompressed. Early results from the pilot, which covered records held at the Stadthaus Zürich on Stadthausquai, indicated that automated tools could flag candidate duplicates at a rate that would take human reviewers months to match.

The practical stakes are not abstract. Zurich's housing shortage has pushed planning departments to process permit applications faster, and inaccurate or duplicated images in a case file can trigger review delays. The Stadtentwicklung Zürich, which coordinates long-term urban planning from its offices near Rathaus, has flagged document-quality issues as a minor but recurring friction point in planning workflows — though no public statement has attributed specific delays to image duplication specifically.

There is also a cost dimension. Swiss cantonal IT budgets are not publicly itemised at the granularity needed to put a precise franc figure on wasted storage, but cloud storage pricing benchmarks suggest that eliminating even a 20 percent duplication rate across a multi-terabyte archive translates into meaningful ongoing savings, particularly as cantons migrate data to hybrid cloud environments under post-2025 procurement rules.

The replacement programme — swapping flagged duplicates with canonical master references — is scheduled to complete its first major phase by the end of 2026. For Zurich residents and businesses navigating permit applications, the practical consequence should be faster document retrieval and more consistent image records in public-facing portals. The Stadtarchiv is expected to publish updated access guidelines for its digitised collections later this year. Institutions waiting on that guidance would be well advised to flag any current discrepancies in document sets before the migration window closes.