How Istanbul's Public Archives Ended Up Drowning in Duplicate Images — and What the City Is Doing About It

Istanbul's municipal digital archive holds more than 4.2 million image files — and a significant portion of them, according to internal assessments cited by the Istanbul Metropolitan Municipality's (İBB) Digital Transformation Directorate, are duplicates. The redundancy problem, years in the making, has finally forced a formal remediation programme that began rolling out across İBB systems in early 2026.

The timing matters. Turkey's broader push toward e-government, accelerated by a national digitalisation strategy adopted in 2021, put every major municipality under pressure to migrate paper and film records into searchable databases. Istanbul, managing assets that range from Ottoman-era cadastral maps to construction permits for the Bosphorus waterfront, moved faster than most — but speed came at a cost. Multiple departments digitised the same holdings independently, often using different resolution standards and file-naming conventions, before any centralised metadata protocol was in place.

How the Duplication Problem Built Up Over Two Decades

The story really starts in the early 2000s. The İBB's Urban Planning Directorate and the separately administered Historic Peninsula Conservation Board both began scanning maps and site photographs around 2003 and 2004, with no shared storage infrastructure. The Atatürk Library in Beyoğlu, which holds one of the country's largest collections of Istanbul photography, launched its own digitisation effort in 2008 under a different technical framework. Each institution retained its own servers, its own naming logic, and its own backup procedures.

By the time the İBB began consolidating these systems under a unified platform around 2019, technicians discovered that thousands of images existed in three or four versions — sometimes scanned from the same original negative at different resolutions, sometimes uploaded multiple times when staff changed and institutional memory was lost. The Kadıköy district archive alone reportedly flagged over 80,000 image entries requiring manual review, according to documentation referenced in a 2024 İBB Digital Transformation progress report.

The 2023 Kahramanmaraş earthquakes added urgency. Istanbul's administration accelerated the digitisation of building inspection records and neighbourhood mapping data as part of earthquake preparedness planning, creating yet another wave of uploads into systems that were already cluttered. Files tied to rapid structural surveys of older residential stock in Fatih and Zeytinburnu — two districts identified as high seismic-risk zones — came in fast and without the quality controls that slower archival work would normally require.

The Cleanup Programme and What It Involves

The current remediation effort, formally launched in February 2026, centres on a deduplication algorithm deployed across the İBB's central storage environment in Başakşehir, where the municipality's main data centre operates. The process uses perceptual hashing — a technique that compares image content rather than just file names or sizes — to flag near-identical images for human review. That distinction matters because many of the duplicates are not exact copies; they are the same photograph scanned at slightly different angles or brightness levels.

The Atatürk Library and the Beyoğlu Cultural Directorate are participating in a pilot phase that runs through September 2026, during which archivists are manually verifying flagged pairs and deciding which version to retain as the canonical record. Librarians there are also building a shared taxonomy that will, for the first time, allow keyword searches to pull results from both the library's holdings and the İBB's planning archive simultaneously.

Costs are real. Dedicated archival storage in a compliant municipal environment runs roughly 0.15 to 0.20 Turkish lira per gigabyte per month for on-premise systems at current procurement rates — a figure that seems small until you multiply it across the terabytes of redundant data sitting idle. Cutting the duplication rate by even 30 percent would free meaningful budget headroom at a moment when the lira's purchasing power continues to constrain municipal IT spending.

For residents and researchers, the practical payoff should arrive by late 2026: a unified public-facing portal, currently in beta testing, that will allow anyone to search Istanbul's digitised visual heritage without knowing which directorate originally held the image. The Atatürk Library's reading room on İstiklal-adjacent Yeniçarşı Caddesi has already posted notices about the transition. Getting here took twenty-odd years of institutional fragmentation. Unwinding it, archivists say, will take at least another eighteen months.