Zurich's Digital Archives Are Riddled With Duplicate Images — and Fixing It Will Reshape How Residents Access Their City's History

Hundreds of thousands of duplicate images have accumulated inside Zurich's civic and institutional digital archives over the past two decades, creating a problem that archivists, researchers and ordinary residents are now being forced to confront together. The Stadt Zürich Stadtarchiv, which holds records stretching back to the medieval period, confirmed earlier this year that a structured deduplication review was underway across its publicly accessible online collections. The scale of the redundancy — estimated internally to affect a significant share of scanned photographic holdings — has slowed search results, confused researchers and, in some cases, led residents to assume that multiple versions of the same image represent distinct historical moments.

This is not a niche technical headache. Zurich's housing shortage has pushed historical planning documents and neighbourhood surveys into daily use, as residents and lawyers comb archives to contest rezoning decisions or support Einsprachen — formal objections — under Swiss planning law. When a search for, say, a 1970s photograph of Langstrasse returns a dozen copies of the same image tagged with slightly different metadata, it wastes time and erodes confidence in the source material. In a city where direct democracy means residents regularly do their own document research before a Volksabstimmung, that erosion carries real civic cost.

Where the Problem Is Felt Most Sharply

The issue is particularly visible at two institutions that thousands of Zurichers use each year. The Zentralbibliothek Zürich on Zähringerplatz — one of the largest research libraries in the German-speaking world — hosts digitised image collections through its e-rara and e-manuscripta platforms. Librarians there have acknowledged the deduplication challenge affects interoperability with partner archives across Switzerland. Meanwhile, the ETH-Bibliothek, part of ETH Zurich on Rämistrasse, maintains its own sprawling image portal, e-pics, which contains more than two million photographs. Duplicate entries in e-pics have at times caused the same image to appear under contradictory captions — an integrity problem the library's digital curation team began formally addressing in 2024.

For residents of Zürich-West or Aussersihl researching the industrial transformation of their neighbourhoods, or for students at the Hochschule der Künste Zürich pulling reference images for thesis projects, duplicate records mean extra filtering work that should not exist. A researcher who retrieves twelve results for a single demolished factory on Hardturmstrasse cannot immediately know whether she is looking at twelve moments or one moment copied eleven times.

What the Clean-Up Means in Practice

Deduplication is not simply deletion. When archivists identify duplicate images, they must decide which version carries the most complete or accurate metadata, merge associated records, and preserve audit trails so that researchers can see what was consolidated and when. The Stadtarchiv's current review is expected to run through the end of 2026. The ETH-Bibliothek has been using automated perceptual hashing — a technique that generates a compact digital fingerprint for each image to detect near-identical copies — since 2023, according to published documentation on the e-pics platform.

The practical consequences for residents will unfold gradually. Cleaner archives mean faster, more reliable searches on platforms that Zurichers already use. Einsprache filings supported by archival photographs will carry more evidential weight when the source record is unambiguous. Teachers at schools across Kreis 3 and Kreis 6 who use digitised maps and historical images in classroom settings will find the material easier to navigate and cite.

The process also has a cost dimension. Swiss federal digitisation funding under the Swissuniversities P-5 program has supported archive modernisation projects, but deduplication work is labour-intensive and often falls outside grant scope, meaning institutions must absorb it from existing budgets. The Zentralbibliothek's annual operating budget runs to tens of millions of francs, yet dedicated digital curation staffing remains limited relative to the volume of holdings.

Residents who use Zurich's civic archives — whether to trace a property boundary, research a family history or prepare for a neighbourhood vote — can expect the search experience to improve noticeably by early 2027, once the current review cycle concludes. In the meantime, archivists at both the Stadtarchiv on Neumarkt and the ETH-Bibliothek recommend noting the specific record identifier, not just the image, when citing digitised material in any formal submission.