How São Paulo's Digital Archives Ended Up Flooded With Duplicate Images — and What Brought Us to This Point

São Paulo's municipal government is sitting on a problem that has been building since at least 2014: its public-facing digital platforms — from the Prefeitura's official portal to the archives managed by the São Paulo Municipal Secretariat of Culture — contain tens of thousands of duplicate images, bloating storage costs, slowing page loads and, in several documented cases, generating legal headaches over licensing rights attached to the wrong version of a file.

The issue matters now because the city is midway through a R$480 million digital transformation programme approved under Mayor Ricardo Nunes in 2024, which is meant to consolidate more than 40 legacy systems into a single integrated platform. Engineers brought in to audit those legacy systems flagged the duplicate-image problem as a structural obstacle — one that predates the current administration and stretches back through at least three successive cycles of outsourced IT contracts.

A History Written in Bad Procurement

The root cause is straightforward, if inglorious. Between roughly 2012 and 2022, the city signed separate technology contracts with different vendors for its health portal, its urban planning database, its tourism promotion site and a cluster of cultural-heritage digitisation projects run out of the Centro Cultural São Paulo on Rua Vergueiro. Each vendor delivered its own content management system. None of those systems spoke to the others. When city staff needed a photograph — say, an aerial shot of Parque Ibirapuera or an image of a flooding event in the Pinheiros River basin — they uploaded it independently to whichever platform they happened to be working in that day. The same photograph, sometimes in different resolutions or with different embedded metadata, exists now across multiple siloed databases.

The Secretaria Municipal de Gestão first formally identified the duplication issue in a 2021 internal audit, according to procurement documents published on the city's transparency portal. That audit estimated the redundant files were consuming storage space equivalent to around 18 terabytes across active servers — a figure that, at the commercial cloud-storage rates the city was paying at the time, translated into wasted expenditure running into hundreds of thousands of reais annually. The audit recommended a deduplication protocol. The recommendation sat largely unimplemented for nearly three years.

Part of the reason for the delay was political. The digitisation contracts were distributed across multiple secretariats, each with its own budget line and its own relationship with its incumbent vendor. Getting agreement across the Secretaria de Cultura, the Secretaria de Saúde and the Secretaria de Desenvolvimento Urbano to adopt a unified asset management approach required a kind of inter-departmental negotiation that the city's bureaucratic structure was not built for. A coordinating committee was established in early 2023 but met irregularly and produced no binding standard before its mandate expired.

The Present Reckoning

The 2024 digital transformation programme changed the calculus. With a single vendor — a consortium led by a São Paulo-based technology firm operating out of the Faria Lima tech corridor — now responsible for migrating data from all legacy systems into one environment, the duplicate images could no longer be quietly ignored. Every file had to be catalogued, and every duplicate had to be either merged, deleted or, in cases where licensing was ambiguous, quarantined pending legal review.

That legal review is not trivial. Some of the images in the municipal archives were contributed by freelance photographers under contracts that specified use on a single named platform. Moving them into a consolidated system — or, worse, discovering that the same image was uploaded under two different contracts — opens questions of copyright compliance that city lawyers are still working through.

For residents and civil society groups that rely on the city's digital services — including the cultural institutions clustered around Avenida Paulista and the neighbourhood associations in Itaim Bibi that use city planning portals to track zoning applications — the practical upshot is simpler: expect slower-than-advertised rollouts of the new unified platform as the cleanup continues. The consortium has set an internal target of completing the deduplication work by the end of the third quarter of 2026, but the legal quarantine backlog means full platform integration is unlikely before early 2027.

For anyone managing digital assets in the public or private sector in São Paulo, the lesson is the same one the city is learning the hard way: without a shared taxonomy, a common repository and mandatory cross-system checks from day one, duplication is not a glitch — it is the inevitable default outcome.