How Paris's Public Image Archives Ended Up Flooded With Duplicates — and What Happened Next

Paris's municipal image libraries contain somewhere in the region of 1.2 million digitised photographs, maps, and architectural drawings spread across more than a dozen institutions — and a significant share of them are the same image filed twice, sometimes three times over. That is the central finding of an internal audit completed in late 2025 by the Direction des Affaires Culturelles de la Ville de Paris, the body that oversees the capital's cultural policy and its associated document repositories.

The timing matters. The city is mid-way through deploying the Grand Paris Numérique framework, a broader digital governance push tied partly to Paris 2024 Olympics legacy commitments. Consolidating visual archives was written into that agenda as a precondition for the open-data portals the city promised to deliver by the end of 2026. Duplicate images clog metadata pipelines, slow search APIs, and inflate cloud storage invoices — problems that are small individually but, at municipal scale, collectively expensive.

A Decade of Well-Intentioned But Disconnected Investment

The problem did not emerge overnight. Between 2014 and 2022, individual institutions — including the Bibliothèque historique de la Ville de Paris on Rue Pavée in the Marais, the Musée Carnavalet on Rue de Sévigné, and the Atelier Parisien d'Urbanisme known as APUR — each digitised their own holdings on separate procurement cycles, using different metadata schemas and incompatible asset-management software. When the city later attempted to aggregate these collections into a unified portal, the ingestion tools had no reliable way to detect that a 1920s photograph of the Pont de la Tournelle already existed under four separate catalogue numbers.

The Carnavalet renovation, completed in 2021 after a four-year closure, accelerated the problem. Digitisation work undertaken during the refurbishment ran in parallel with older scans already held by the Bibliothèque historique, but the two institutions were not comparing notes. By the time collections managers began cross-referencing in early 2024 — prompted partly by preparation for the Olympics cultural programme — they found hundreds of duplicate entries for iconic imagery of Haussmann-era boulevards and early twentieth-century street life in arrondissements including the 11th and 13th.

The Audit and What Comes Next

The 2025 Direction des Affaires Culturelles audit reportedly identified over 80,000 records as probable or confirmed duplicates across the consolidated digital estate, though the figure covers only the institutions directly under city jurisdiction and excludes bodies such as the BnF, which operates under national rather than municipal authority. Resolving each record requires human review: automated hash-matching catches identical files but misses cases where the same image was scanned twice at different resolutions or with different cropping.

The financial case for fixing this is straightforward. Cloud storage costs for Paris's cultural directorate have risen steadily since 2020 as digitisation volumes grew. Deduplication projects in comparable European municipal archives — notably Amsterdam's Stadsarchief, which completed a similar exercise in 2023 — have reported storage reductions of between 15 and 22 percent following systematic duplicate removal. Paris has not published equivalent projections for its own estate.

The city has contracted a phased clean-up running through the first quarter of 2027, timed to align with the expected launch of the unified open-data image portal tied to Grand Paris Numérique commitments. Collections managers at the Carnavalet and the Bibliothèque historique are understood to be piloting a shared metadata standard based on the Dublin Core schema, which would prevent new duplicates from accumulating as fresh digitisation work continues along the Seine regeneration corridor and in the banlieue-focused heritage programmes attached to suburban Grand Paris Express station openings. Whether the timeline holds depends on staffing levels that have not yet been confirmed in the 2026-27 municipal budget cycle — a budget currently under pressure as the National Assembly scrutinises city-level spending across the Île-de-France region.