Zurich Archives Push to Fix Duplicate Image Crisis Across City Digital Collections

City archivists and digital librarians in Zurich confirmed this week that a multi-institution sweep to clean up duplicate image files in public collections is moving into its most intensive phase yet, after months of preparatory work revealed the scale of the problem runs deeper than initially expected.

The effort matters now because the UBS cultural heritage digitisation grants — part of a broader civic tech programme tied to the bank's post-Credit Suisse restructuring commitments — have set a December 2026 deadline for recipient institutions to demonstrate clean, audit-ready digital catalogues. Duplicates inflate storage costs, distort search results, and can trigger copyright re-licensing complications when the same photograph appears under different metadata entries. With Zurich's housing shortage already straining municipal budgets, institutions cannot afford redundant infrastructure spending.

What Happened This Week

Staff at the Stadtarchiv Zürich on Alfred-Escher-Strasse and the Zentralbibliothek Zürich on Zähringerplatz began coordinating directly on Wednesday, using a perceptual-hashing algorithm to flag visually identical or near-identical images stored under separate accession numbers. The technique compares compressed image fingerprints rather than pixel-by-pixel data, allowing it to catch duplicates even when one copy has been slightly cropped or re-scanned at a different resolution.

ETH Zürich's Data and Service Center for the Humanities, known as DaSCH, is providing technical support. The centre, based on Rämistrasse, has worked on similar deduplication projects for Swiss humanities institutions since at least 2021 and maintains infrastructure used by research teams across the German-speaking world. DaSCH's involvement is significant: its tools are already integrated with the Swiss national metadata standard, meaning any verified clean records can feed directly into the national IIIF image repository without further reformatting.

Separately, the Museum Rietberg in Rieterpark flagged an internal audit completed on 30 June that identified just over 1,400 image entries in its online collection portal carrying potential duplicate status — roughly 4.2 percent of its publicly visible digital holdings. The museum has not yet stated how many of those will be removed versus merged, but the audit itself is the first publicly confirmed dataset to emerge from this week's activity and gives the broader city effort a concrete benchmark to work against.

Why This Is More Than a Housekeeping Exercise

The practical stakes are higher than they might appear. Zurich Cantonal law on public records, updated in January 2025, now requires that digitised public assets be findable and uniquely identifiable within cantonal systems by the end of the current budget cycle. Duplicate entries technically violate that requirement if they cause the same object to appear as two distinct records. Institutions that cannot demonstrate compliance risk losing access to the digitisation grant tranche scheduled for disbursement in the first quarter of 2027.

Storage is also a live cost issue. The city's central IT procurement framework — administered through Organisation und Informatik, the municipal IT department — renews cloud storage contracts on a rolling 18-month basis. The next renewal window opens in September 2026. Cutting confirmed duplicates before then could reduce contracted capacity requirements and, by extension, lower the figures that land in next year's budget submission to the Stadtrat.

For researchers and members of the public who rely on platforms like e-rara.ch or the Zürich cantonal image portal, deduplication also means cleaner search results. A photograph that currently appears three times under three different catalogue numbers collapses into a single, properly attributed entry — relevant for anyone tracing historical images of, say, Langstrasse's pre-war streetscape or construction records from the Zürichberg development projects of the 1960s.

The institutions involved expect a first round of confirmed deletions and merges to be completed by late August. Any member of the public who has bookmarked or cited specific image URLs from the Stadtarchiv or Zentralbibliothek collections should check whether those links remain valid after the consolidation, as canonical URLs for merged records will change. Both institutions say they plan to publish redirect tables when the process is complete.