Zurich's Digital Archives Are Full of Duplicate Images — and Officials Say the Cleanup Bill Is Rising

Zurich's public institutions are sitting on millions of redundant digital image files, and the people responsible for managing those archives are no longer willing to stay quiet about it. Curators at the Stadtarchiv Zürich on Alfred-Escher-Strasse and technology officers at ETH Zurich's library system have separately flagged the issue in recent months, pointing to a structural problem that has quietly inflated storage budgets and complicated public access to historical records.

Duplicate image replacement — the process of identifying, consolidating and removing redundant copies of digitised photographs, scans and graphics across institutional databases — has become a pressing administrative concern across European municipalities. In Zurich, where the digitisation of public records accelerated sharply after 2020, the problem has reached a scale that specialists describe as no longer manageable through manual review alone.

Why It Matters Now

The timing is not accidental. Zurich's city administration has been expanding its digital infrastructure under the Smart City Zurich programme, which has set milestones for open-data availability running through 2027. That push has drawn new attention to the underlying quality of what is actually in those systems. When the same scanned photograph of, say, the old Hauptbahnhof concourse exists in fourteen slightly different file versions across three separate databases, every downstream application — from public search tools to AI-assisted cataloguing — inherits that mess.

Storage is not cheap at institutional scale. Cloud and on-premises data management for mid-sized European city administrations typically runs into the tens of millions of francs over a multi-year contract cycle, and a significant fraction of that cost is attributable to redundant assets. Zurich Stadtwerk and the city's IT department have not published a specific figure for duplicated image overhead, but technology procurement specialists working in the Swiss public sector have pointed to duplication rates of 20 to 40 percent as common in archives that digitised rapidly without a unified metadata standard.

ETH Zurich's Scientific IT Services group, based on the Hönggerberg campus, has been developing automated deduplication tools as part of broader research data management work. The university handles an enormous volume of scientific imagery — microscopy outputs, satellite imagery, experimental data — and the challenge of distinguishing a true duplicate from a legitimate variant of a similar image is technically non-trivial. Perceptual hashing algorithms and content-based retrieval systems are among the approaches being tested, though researchers caution that no single method works cleanly across all image types.

What the Professionals Are Saying

Heritage and information professionals in Switzerland have grown more vocal. The Verein Schweizerischer Archivarinnen und Archivare — the national archivists' association — held a working session on digital asset governance in Bern in May 2026, with participants from cantonal and municipal institutions across the German-speaking regions. The core message from that gathering, according to published meeting notes, was that deduplication cannot be treated as a one-time technical fix; it requires ongoing governance policy baked into digitisation workflows from the start.

At the Zentralbibliothek Zürich on Zähringerplatz, which holds one of the largest collections of historical Swiss photographs and printed matter, staff have been working since early 2025 to reconcile image metadata across its e-manuscripta and e-rara platforms. The library has not disclosed how many duplicate records have been identified, but the project is understood to be ongoing and resource-intensive.

For smaller institutions — neighbourhood cultural centres, district archives in areas like Wiedikon or Oerlikon, local historical societies — the technical and financial burden is proportionally heavier. They rarely have dedicated IT staff, and off-the-shelf deduplication software licences can run from several hundred to several thousand francs annually depending on the volume of assets managed.

The practical advice from specialists is consistent: establish a single authoritative master file standard before migrating images between systems, assign persistent identifiers to every asset at the point of ingest, and schedule regular automated audits rather than waiting for storage costs to force a reckoning. For institutions already deep in the problem, a phased approach — prioritising high-traffic public collections first — is considered more realistic than attempting a full-database sweep in one pass. The Smart City Zurich programme's next reporting cycle is due in the first quarter of 2027, and data quality metrics are expected to feature prominently in that review.