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

Stockholm's municipal archive contains tens of thousands of duplicate photographs. That is the core finding driving a city-wide initiative launched earlier this year to rationalise how the capital stores, tags and shares visual records across its sprawling network of administrative bodies. The problem did not appear overnight.

The story of how Stockholm arrived here runs back roughly fifteen years, to the period between 2010 and 2014 when individual district councils — from Södermalm to Östermalm — began digitising their own holdings independently, without a shared technical standard. Each district bought its own content management software, hired its own contractors, and ingested photographs at whatever resolution and naming convention seemed locally sensible at the time. The result was predictable in retrospect: the same image of, say, the Katarina Kyrka restoration or a Midsommar event on Djurgården could exist in six separate databases under six different file names, each with slightly different metadata.

The Bureaucratic Roots of a Digital Mess

Stockholms Stadsarkiv, the city's official records office on Kungsklippan, spent much of the 2010s focused on physical document preservation. Digital image governance was not a stated priority in the office's formal remit until a review conducted by the Swedish National Archives — Riksarkivet — flagged the issue in a 2019 sector-wide assessment. That review found that municipal bodies across Sweden were collectively spending an estimated 12 percent of their digital storage budgets on redundant file copies, a figure that translated into millions of kronor annually when scaled to a city the size of Stockholm.

The pandemic years between 2020 and 2022 complicated matters further. Remote working pushed city employees toward ad-hoc cloud storage solutions. Staff at Stadsbyggnadskontoret, the city planning office on Fleminggatan, began uploading photographs to shared drives that sat outside the official archival system entirely. When those projects wrapped up, the images were rarely migrated back into Stadsarkivet's infrastructure with proper cataloguing. Some were duplicated again simply because colleagues could not locate the originals.

A pilot consolidation project run through Kulturförvaltningen — Stockholm's cultural administration — during 2023 attempted to merge records from three inner-city district databases. The exercise took eight months and cost approximately 1.4 million kronor according to budget documents from that fiscal year. It revealed that around 34 percent of images in the test sample were functional duplicates, meaning identical or near-identical photographs that had been stored separately with no cross-reference between them.

The Push for a Unified System

City council approved a broader remediation programme in February 2026, allocating funds through the digital infrastructure budget for a phased duplicate-detection rollout across all fourteen district councils. The programme uses algorithmic image-matching tools already deployed by institutions including the Nordiska museet on Djurgårdsvägen, which completed its own internal deduplication exercise in late 2024. The city is working with that institution's technical team to adapt the approach for administrative rather than cultural collections.

The timeline is tight. Phase one, covering the inner districts including Kungsholmen and Vasastan, is scheduled for completion by the end of October 2026. Phase two, addressing the outer suburbs, runs through the first quarter of 2027. Officials at Stadsarkivet have acknowledged publicly — in written responses to council questions — that the October deadline assumes no significant staffing disruptions.

For residents and journalists who rely on the city's public image library for planning documents, heritage records, or neighbourhood histories, the practical upshot is a period of partial access to some collections. Stadsarkivet advises users who need images from affected databases to contact the archive directly at Kungsklippan and submit a formal retrieval request, which should be processed within ten working days under the current service standard. Once the deduplication is complete, the city promises a single searchable portal — replacing the current patchwork of district-level databases — where all consolidated records will be publicly accessible under Creative Commons licensing terms.

Whether the October target holds will depend partly on how many further duplicates the algorithmic scan turns up. The 2023 pilot suggested the scale of redundancy surprised even the project managers. The full archive is considerably larger.