Zurich Archives Overhaul: Duplicate Image Replacement Effort Reaches Critical Phase This Week

Zurich's municipal digitisation programme crossed a threshold this week when the Stadtarchiv Zürich confirmed that automated duplicate-image replacement across its public-facing holdings had cleared more than 40,000 redundant file entries — a figure that represents roughly one-fifth of the backlog identified when the project launched in January 2025. The work, concentrated on the archive's Neumarkt facility in the Altstadt district, is being watched closely by Swiss federal records bodies as a potential model for cantonal adoption.

The timing matters. Switzerland's revised Federal Act on Data Protection, which came into force in September 2023, placed new obligations on public institutions to maintain clean, non-duplicated personal data records. Images fall squarely within that framework. For Zurich specifically, the pressure is compounded by the sheer volume of digitised material accumulated during the Covid-era push to make civic records remotely accessible — a drive that was fast but not always clean, leaving thousands of near-identical scans sitting across multiple servers.

How the Deduplication Pipeline Actually Works

The technical backbone comes from ETH Zurich's Computer Vision Lab, whose perceptual hashing tools are being used under a contract with the city administration. Rather than simple file-name matching, the system compares image fingerprints — allowing it to flag scans where the content is identical even if resolution, format or metadata differ. That distinction is what makes it useful for archival work, where the same photograph may have been digitised at different points in time by different departments.

The Stadtarchiv is not working alone. The Zentralbibliothek Zürich on Zähringerplatz, which holds one of the largest photographic collections in German-speaking Europe, began a parallel deduplication exercise in March 2026 covering its postcard and press-photography holdings. Librarians there have reportedly flagged the process as labour-intensive at the review stage: automated tools can identify probable duplicates, but human archivists must still sign off before any file is retired or replaced with a canonical master copy.

Cost is a real factor. Municipal budget documents published in spring 2026 allocated CHF 2.3 million to the broader digitisation quality programme for the 2026 fiscal year, of which duplicate-image remediation is one workstream among several. That allocation runs through December 2026, meaning the current pace must be sustained — or accelerated — if project leads want the core backlog cleared before funding cycles reset.

Why Ordinary Zürich Residents Should Care

The practical stakes extend beyond archivists. The city's online planning portal, used by residents in districts like Wiedikon and Oerlikon to review building permit applications and neighbourhood planning documents, draws on the same image repositories. Duplicate or mislinked images have caused at least three documented instances in 2025 where the wrong site photograph appeared against a planning application — errors that generated formal objections and delayed decisions by weeks, according to published minutes of Stadtrat working sessions available through the city website.

The Swiss direct-democracy system amplifies such glitches. When residents preparing for a cantonal referendum or a communal Abstimmung consult digital records to inform their vote, data quality is not a bureaucratic nicety — it is a precondition for informed civic participation. Any photograph mislabelled or duplicated in the official record can muddy the evidentiary basis of a neighbourhood objection or a planning appeal.

Project managers at the Stadtarchiv are expected to publish an interim progress report by the end of July 2026, which will include updated counts on files processed, error rates at the automated-flagging stage, and a revised estimate for full completion. Institutions elsewhere in the canton — including Winterthur's Stadtbibliothek — have already signalled interest in licensing the ETH-developed toolset. Residents or organisations wishing to flag specific records they believe contain duplicate or mismatched images can submit requests through the Stadtarchiv's online portal at stadtarchiv.stadt-zuerich.ch, where a dedicated deduplication-query form went live on 1 July 2026.