Paris's urban planning authority, the Atelier Parisien d'Urbanisme (APUR), confirmed this spring that it had identified more than 340,000 duplicate or misattributed images inside the city's official photographic and cartographic databases — records that feed everything from tourist information platforms to Grand Paris Express construction planning tools. The figure, circulated internally during a March 2026 review of the city's open-data holdings, underscores a problem that municipal archivists have been flagging since at least 2022 but that only gained political urgency after the Paris 2024 Olympics placed the capital's digital infrastructure under unprecedented scrutiny.
The stakes are higher than they might appear. Municipal databases that carry duplicate or mislabelled images don't just inconvenience archivists — they distort planning decisions, create legal liability around image rights, and corrode public trust in open-data portals that residents are increasingly asked to rely on. With the Grand Paris Express metro expansion still active across a dozen construction sites from Saint-Denis to Bagneux, accurate geospatial imagery is not an administrative nicety. It is a logistical requirement.
What Paris Is Actually Doing
The Ville de Paris launched a dedicated deduplication programme in January 2026 under the umbrella of its Direction des Systèmes et Technologies de l'Information (DSTI). The unit is working with open-source image-hashing tools and a proprietary verification layer developed partly in partnership with the Conservatoire National des Arts et Métiers, whose Rue Saint-Martin campus has become an informal hub for the city's civic-tech experimentation. The target, according to programme documentation published on the city's open-data portal in February, is to reduce confirmed duplicates in the municipal image registry by 60 percent before the end of 2026.
The effort is concentrated in two areas. First, the legacy photographic archive of Paris Musées — the network of 14 municipal museums including the Musée Carnavalet on Rue des Francs-Bourgeois — which digitised roughly 150,000 images between 2018 and 2023, a process that introduced a significant rate of duplicate entries when collections were merged. Second, the Seine-Saint-Denis prefecture's shared planning image library, which feeds into both suburban housing applications and the ongoing Seine riverside regeneration corridor between Bercy and Ivry-sur-Seine.
How Paris Compares Globally
Other major cities have moved earlier and, in some cases, more decisively. Amsterdam's municipal archive, Stadsarchief Amsterdam, completed a full deduplication sweep of its 800,000-image digital collection in 2024, deploying a machine-learning pipeline built in collaboration with the University of Amsterdam. The Dutch capital reduced duplicate records by an estimated 72 percent over 18 months. Tokyo's National Diet Library, which manages image metadata for dozens of ward-level municipalities, introduced mandatory hash-verification for all incoming digitised materials in 2023, effectively preventing the problem from accumulating in the first place.
London's situation is arguably messier than Paris's. The Greater London Authority does not operate a single unified image database — records are fragmented across 32 borough councils, Transport for London, and Historic England — making any city-wide deduplication exercise largely theoretical. By that comparison, Paris's centralised approach through APUR and DSTI gives it a structural advantage that it has, until recently, been slow to exploit.
New York City's Department of City Planning completed a deduplication audit of its BYTES of the BIG APPLE open-data platform in late 2024 and publicly reported a 41 percent reduction in duplicate geospatial image files. Paris's 60 percent target, if met, would represent a stronger outcome — though the French capital is starting the formal programme roughly a year behind New York's timeline.
For Parisians and professionals who rely on the city's open-data ecosystem, the practical advice is straightforward for now: cross-reference any image or geospatial file pulled from paris.fr or data.gouv.fr against at least one secondary source before using it in planning, legal, or journalistic work. APUR has published a verification checklist on its website, updated as of April 2026, that details which dataset categories remain under active deduplication review. The Seine riverside corridor files and the 10th and 19th arrondissement planning archives are specifically flagged as incomplete. A full progress report is expected from DSTI before the end of September.