Zurich's Digital Archives Are Drowning in Duplicate Images — And the Numbers Tell a Costly Story

Duplicate images cost Swiss public institutions an estimated 12 to 18 percent of their total digital storage expenditure each year, according to data compiled by the Swiss Federal Archives in Bern — and Zurich, as the country's most data-intensive urban centre, is disproportionately exposed. The problem sits at the intersection of municipal bureaucracy, academic research infrastructure, and a housing-shortage crisis that has driven an explosion of property listing photographs across cantonal planning databases.

The timing matters. Zurich's city administration is currently mid-way through its Digitale Stadt Zürich programme, a multi-year initiative to migrate legacy document systems onto unified cloud infrastructure. That migration, which began in 2023 and is scheduled to run through late 2027, has forced IT departments across Stadthaus Zürich and the Amt für Städtebau — the urban planning office on Lindenhofstrasse — to confront just how badly duplicate files have fragmented their archives. Planning files for contested rezoning applications in districts like Altstetten and Schwamendingen frequently contain the same site photographs uploaded four, five, or six times across different departmental folders, each copy consuming server space at a rate that adds up.

What the Storage Numbers Actually Show

ETH Zurich's central IT services division reported in its 2025 annual infrastructure review that roughly 23 percent of files stored across its research data repositories were exact or near-exact duplicates. For a university that manages over 4.8 petabytes of active research data — a figure cited in the university's own published infrastructure documents — that represents close to 1.1 petabytes of theoretically redundant storage. At current colocation and cloud pricing in the Swiss market, where enterprise storage runs between CHF 0.08 and CHF 0.14 per gigabyte per month depending on redundancy tier, the annual carrying cost of that duplication alone reaches into seven figures.

The Zentralbibliothek Zürich on Zähringerplatz faces a parallel challenge in its digitisation programme. The library has been scanning historical maps, photographs, and manuscripts since 2018 as part of its e-rara and e-manuscripta partnerships. Automated deduplication tools catch obvious copies, but near-duplicate images — a scan taken twice at marginally different exposures, or the same photograph ingested from two separate donor collections — slip through standard hash-matching algorithms and require more computationally expensive perceptual hashing techniques that the library has only recently begun piloting.

Why the Housing Crisis Made It Worse

Zurich's Wohnungsnot crisis has added an unexpected layer to the problem. The cantonal building permit database and the city's own Immobilienamt have seen a sharp rise in submitted documentation since 2022, as developers rushed conversion and densification applications ahead of tightening zoning rules. Each application can attach dozens of site images. When applicants resubmit amended plans — common in a system where direct democracy means neighbourhood referendums can force redesigns — the original image sets persist alongside the new ones. The Amt für Städtebau told a 2025 cantonal audit committee that its document management system held approximately 340,000 image files related to building applications, but had not completed a deduplication audit since 2019.

Commercial platforms have moved faster than government. Homegate, the property listing portal headquartered in Schlieren just west of Zurich, implemented automated duplicate-image detection across its Swiss listings database in early 2024, citing measurable reductions in load times and a cleaner user experience for the roughly 2.1 million monthly unique visitors its Swiss platform attracts, according to figures the company published in its 2024 media kit.

For institutions still working through the problem, the practical path forward involves three steps: running a full perceptual hash audit to identify near-duplicates that byte-level matching misses, establishing a clear retention policy before deletion to avoid destroying legally required originals, and integrating deduplication checks into upload workflows rather than treating it as a clean-up task. The Digitale Stadt Zürich programme office is expected to publish updated file governance standards for all city departments by the fourth quarter of 2026. Whether those standards will include mandatory deduplication thresholds is still being negotiated between the Stadtpräsidium and individual departmental IT leads.