Berlin Takes On the Duplicate Image Problem: How the City Stacks Up Against Amsterdam, Vienna and Seoul

Berlin's Senate Department for Urban Development quietly acknowledged this spring that its publicly accessible planning portal, FIS-Broker, contains thousands of duplicate image files — scanned maps, aerial photographs and site-plan attachments uploaded multiple times by different district offices, sometimes dozens of copies of a single document. The problem has slowed the portal's search function and complicated efforts by housing developers, architects and researchers who rely on it to navigate the city's notoriously complex permitting landscape.

The issue matters now because Berlin is in the middle of a €45 million digitisation push, the Berliner Digitalisierungsprogramm, which the SPD-led Senate launched in 2024 with a 2028 completion target. Duplicate image data inflates storage costs, creates version-control confusion and undermines the credibility of the digital-first promise the city made to residents and businesses. With the housing shortage squeezing Prenzlauer Berg and Neukölln alike, any friction in the planning system carries a real cost.

What Berlin Is Actually Doing

The Senatsverwaltung für Stadtentwicklung, Bauen und Wohnen has partnered with the Zuse Institut Berlin — the state-affiliated research institute on Takustraße in Dahlem — to pilot a deduplication algorithm across a subset of the FIS-Broker image library. The pilot, which began in January 2026, targets georeferenced raster images first, the category most prone to accidental re-upload by staff in the city's 12 autonomous borough administrations. Tempelhof-Schöneberg and Mitte are the two districts participating in the first phase. A broader rollout across all boroughs is scheduled for the fourth quarter of 2026, though officials have not committed publicly to a firm date.

The Zuse Institut is not working alone. The city has also brought in the open-source community around the Wikimedia Foundation's GLAM (Galleries, Libraries, Archives and Museums) initiative, which has helped cultural institutions on Unter den Linden and at the Stadtmuseum Berlin identify and retire redundant image assets in public collections. That crossover between civic tech and cultural heritage is unusual, and it reflects Berlin's startup-friendly instinct to blur institutional lines — a useful habit, even if it sometimes produces coordination headaches.

How Berlin Compares to Amsterdam, Vienna and Seoul

Amsterdam's Gemeente Amsterdam began a systematic deduplication programme for its public geodata portal in 2022, contracting the work to a Dutch civic-tech firm and completing the first full sweep by early 2024. The result, according to the city's published open-data report, was a 34 percent reduction in redundant image files across its spatial planning database. Berlin, which has a comparable volume of legacy scanned documents, is starting from a similar baseline but roughly four years behind Amsterdam's timeline.

Vienna's Stadt Wien took a different route. Rather than running a retrospective cleanup, it redesigned its Wiener Stadtplan upload workflow in 2021 to include automatic hash-matching — a standard technique that flags identical files before they enter the archive. New duplicates have been effectively blocked at the point of entry since then. Berlin's legacy system on FIS-Broker predates that kind of embedded check, which is why the city is stuck doing the harder, retrospective work now.

Seoul's Smart City platform, managed by the Seoul Digital Foundation, applied machine-learning image fingerprinting to its urban data lake in 2023, covering aerial photography going back to 2005. The South Korean capital processed roughly 2.8 million images in that project. Berlin's archive is smaller, but the principle — fingerprint first, deduplicate second, then prevent recurrence — is the same approach the Zuse Institut pilot is trying to replicate on Takustraße.

For Berlin residents and professionals trying to use FIS-Broker right now, the practical advice is straightforward: when downloading planning documents, check the upload timestamp and the originating district office listed in the file metadata. Duplicate files often carry different district tags on identical content. The Senate department has published a one-page guidance note on its website advising users how to cross-reference the document reference number — the Vorgangsnummer — against the master register to confirm they have the authoritative version. The deduplication pilot is expected to produce an interim report by September 2026, which will determine whether the fourth-quarter rollout proceeds on schedule or slips into 2027.