São Paulo Moves to Purge Duplicate Images From City Records — But Lags Behind London and Seoul

São Paulo's city hall has been quietly battling a problem that sounds mundane but carries real fiscal weight: tens of thousands of duplicate images clogging the municipal digital archive, slowing public-records searches and driving up server costs at a time when the Ricardo Nunes administration is already under budget pressure. The issue spans everything from urban planning permits filed through the Secretaria Municipal de Urbanismo e Licenciamento to health-clinic records managed by the Secretaria Municipal de Saúde across the city's 96 subprefeituras.

The problem has sharpened in 2026 because a federal directive issued in January by the Ministério da Gestão e da Inovação em Serviços Públicos requires all municipal administrations with populations above one million to present digital-asset audits by December 31. São Paulo — home to roughly 12.3 million residents according to IBGE's 2022 census estimate — sits squarely in scope. Non-compliance risks the suspension of federal technology-modernisation transfers, a line item that reached R$ 47 million in last year's municipal budget cycle.

What the City Is Actually Doing

The Empresa de Tecnologia da Informação e Comunicação do Município de São Paulo, known as PRODAM, is running the deduplication drive. PRODAM began a phased audit in March, starting with image files tied to infrastructure projects along Avenida Paulista and the Marginal Pinheiros corridor — two zones with the heaviest permit-filing volumes in the city. The agency is using open-source hash-matching software to flag identical or near-identical files before a human review team makes deletion decisions. No proprietary vendor contract has been publicly disclosed for the current phase.

Pinheiros and the Mooca district were chosen as pilot neighbourhoods partly because both saw heavy digitisation pushes during pandemic-era permit freezes in 2020 and 2021, when scanning backlogs meant the same document was often uploaded multiple times by different clerks. PRODAM has not yet published a completion rate for the pilot, though the agency's published work schedule, available on the city's Diário Oficial, lists a mid-August checkpoint for the first progress report.

São Paulo's Centro de Operações — the command hub at Rua Líbero Badaró — is also involved, since it ingests thousands of CCTV and drone images daily. Duplicate frames from flood-monitoring cameras along the Córrego do Mandaqui and Córrego Ipiranga drainage systems have historically inflated storage demands during the November–March storm season. City engineers say the flood-monitoring archive alone had grown to an unmanageable size by early 2025, though no official figure has been made public.

How São Paulo Stacks Up Globally

London's Government Digital Service completed a comparable deduplication project for the Greater London Authority in 2023, reducing its image repository by roughly 34 percent and cutting annual cloud-storage costs by £1.2 million, according to a GLA published report. Seoul's Smart City Division finished a similar exercise for its public CCTV archive in late 2024, processing more than 800 terabytes of footage using AI-assisted deduplication tools developed in partnership with KAIST, the Korea Advanced Institute of Science and Technology. Both cities had a structural advantage São Paulo lacks: a single unified IT governance body with authority over all municipal departments, rather than the fragmented secretaria-by-secretaria structure that PRODAM must negotiate.

New York City's Department of Records and Information Services tackled document-image deduplication as part of its NYC Digital Archives Project, but that effort focused on historical scanned records rather than live operational data — a narrower scope than what São Paulo is attempting simultaneously across active city services.

The cost comparison is striking. London spent roughly £2.3 million on its 2023 project. São Paulo has not published a budget for the PRODAM initiative, but municipal IT consultants familiar with similar Brazilian projects estimate that a city of São Paulo's scale should expect to spend between R$ 8 million and R$ 15 million to complete a full-cycle deduplication across all secretarias — a range consistent with comparable state-level projects in Rio de Janeiro and Minas Gerais over the past three years.

For residents and businesses, the practical payoff is faster response times when pulling building permits or accessing public health records through the SP156 platform. PRODAM's August checkpoint will be the first real indicator of whether São Paulo can meet the federal December deadline — and whether it can genuinely close the gap on cities that started this work years earlier.