How Zurich's Digital Archives Ended Up Full of Duplicate Images — and What's Being Done About It

Zurich's municipal digital archive holds hundreds of thousands of images. A growing share of them are duplicates. That is the finding driving a coordinated effort across several city departments to clean, deduplicate, and restructure how photographs enter and circulate through public databases — a problem that has compounded quietly since at least 2018, when the city accelerated its push toward paperless administration under the Smart City Zürich programme.

The issue is not trivial. Redundant images inflate storage costs, slow retrieval systems used by city planners and journalists alike, and — most critically for institutions bound by Swiss data protection law — create compliance risks when the same identifiable photograph of a person or location sits in three separate repositories under different file names. The Federal Act on Data Protection, revised most recently in September 2023, tightened obligations around exactly this kind of unintended data retention.

How the Problem Accumulated

The roots go back further than 2018. Zurich's transition to digital record-keeping gathered pace through the 2010s, but different departments built their own image libraries in isolation. Stadtentwicklung Zürich, the city's urban development office headquartered near Stadthaus on Stadthausquai, ran its own media library. So did the communications team at the Rathaus on Limmatquai. When departments collaborated on projects — say, the ongoing Hochhausstrategie debates around permissible tower heights in Zürich-West — images were shared by email, downloaded, renamed, and re-uploaded. Each step created a new file instance.

ETH Zurich's chair of information science has studied this pattern in Swiss public-sector organisations more broadly. Research published through the university's main campus on Rämistrasse has documented how siloed digital migration, conducted department by department rather than city-wide, reliably produces duplication rates of between 15 and 30 percent in municipal image databases within five years of digitalisation. Zurich's own internal audits, described in publicly available Stadtrat session summaries from late 2024, acknowledged the problem without publishing specific figures.

The post-merger restructuring at UBS, following its absorption of Credit Suisse in 2023, added an unexpected dimension. Several private-sector documentation platforms used by Zurich-area financial institutions were renegotiated or discontinued, pushing some archive-management vendors to refocus on public-sector contracts. That shifted pricing and product availability for the tools Zurich's institutions rely on, forcing a reassessment of long-term digital infrastructure that brought the duplication problem into sharper focus.

What Comes Next

The city is now piloting an automated deduplication protocol across a subset of departments. The pilot, running through the end of 2026, uses perceptual hashing — a technique that identifies visually identical or near-identical images regardless of file name or format — to flag redundant entries before a human archivist makes a final call. The Stadtarchiv Zürich on Neumarkt is understood to be among the participating institutions, though a formal public announcement has not yet been made.

For ordinary Zurichers, the immediate effect is modest. The housing shortage on platforms like the city's official rental portal means photographs of available Wohnungen are among the most viewed public images the city hosts; better-organised archives could improve the accuracy and speed of those listings. More broadly, researchers at institutions along Rämistrasse and journalists filing public records requests stand to benefit from faster, cleaner search results.

The practical advice for anyone submitting images to city systems right now is straightforward: use standardised file naming conventions tied to project codes, submit through the single designated intake portal rather than by email, and retain original high-resolution files locally before upload. Those habits will not fix the inherited backlog, but they will stop the duplication rate from climbing further while the cleanup runs its course.

The pilot's results are expected to inform a city-wide policy recommendation before the Stadtrat by spring 2027. Whether that recommendation will include a mandatory centralised image management platform — or simply tighter procedural rules around existing tools — remains an open question. The audit findings due later this year should narrow that down considerably.