How Zurich's Digital Archives Fell Into the Duplicate Image Problem — and What It Took to Get Here

Zurich's public institutions are sitting on millions of redundant digital image files. That is not a guess — archivists at the Stadtarchiv Zürich, the city's official records repository on Alfred-Escher-Strasse, have been working since early 2024 to quantify the scale of a problem that built up quietly over roughly two decades of uncoordinated digitisation drives. The core issue is straightforward: when different departments scan the same photograph, document page or map at different times, often using different equipment and naming conventions, the result is a warehouse of duplicates that consumes storage, slows retrieval and erodes the integrity of the public record.

The timing matters. Zurich's ongoing Wohnungsnot crisis has pushed city planners to comb historical building records at an unprecedented rate, looking for precedents and legal baselines for densification projects in districts like Altstetten and Schwamendingen. When the same cadastral photograph exists under four different filenames across three separate departmental servers, a planning officer can waste hours cross-referencing what should be a five-minute lookup. That friction has given the duplicate image problem a new urgency it lacked when the archives were mainly the concern of historians.

How the Backlog Built Up

The roots go back to the early 2000s. Swiss federal digitisation mandates pushed cantonal and municipal bodies to move paper records online, but no single interoperability standard governed how Zurich's roughly 30 departments handled image metadata. The Bauarchiv, which holds construction permits and site photographs, developed its own file-naming logic. So did the Amt für Städtebau, located near Lindenhof, and the Stadtbibliothek Zürich network. By the time ETH Zürich researchers published a 2019 study on public-sector data quality in Swiss cities — citing metadata inconsistency as a primary cost driver — the layering had already produced what one published analysis described as a duplication rate exceeding 30 percent in some municipal image collections.

The UBS-Credit Suisse merger aftermath added an unexpected dimension. Several thousand historical banking and commercial property photographs, previously held in private institutional archives, transferred into semi-public custody between 2023 and 2025 as part of asset rationalisation processes. Some of those images overlapped with material already held by the Stadtarchiv, creating a fresh wave of duplicates that had nothing to do with public-sector scanning practices. The Stadtarchiv confirmed in its 2025 annual report that it received image batches from at least two legacy Credit Suisse real-estate subsidiaries, though the precise volume has not been publicly disclosed.

What the Reckoning Looks Like

The practical response has been methodical rather than dramatic. Since January 2025, the Stadtarchiv has been piloting a deduplication workflow using perceptual hashing — a technique that identifies visually identical or near-identical images even when file names and formats differ. The pilot covered roughly 180,000 images drawn from the Hochbaudepartement's pre-2010 holdings. Early results, shared at a February 2026 archivists' conference in Bern, suggested the process flagged approximately 22 percent of that subset as confirmed or probable duplicates.

The canton's Staatskanzlei, which coordinates cross-departmental IT policy from its offices near the Rathausbrücke, has been drafting a unified image metadata standard since mid-2025. Under the proposed framework, every image entering a public archive would receive a canonical identifier tied to a central registry — preventing the same photograph from being registered independently by two departments. A consultation round with affected agencies closed in May 2026, and a formal rollout is expected before the end of the year, subject to cantonal council approval.

For residents and researchers using the Stadtarchiv's public reading room or its online portal, the practical advice is simple: if you find what appear to be duplicate results in a search, report them using the feedback function on the archive's website. The Stadtarchiv has said it is actively using those reports to prioritise which collections to process next. The work is slow, the backlog is large, but the map of how Zurich arrived at this point is now, at least, clearly drawn.