São Paulo's Duplicate Image Crisis: The Key Decisions That Will Shape the City's Digital Archive

São Paulo's municipal digital archive holds millions of photographs — surveillance footage stills, urban planning records, social programme documentation, infrastructure inspection images — and a growing share of them are duplicates. City Hall acknowledged the problem formally in June 2026 when the Secretaria Municipal de Inovação e Tecnologia flagged that redundant image files were consuming an estimated 30 percent of allocated cloud storage across several departments, driving up operational costs and complicating data retrieval for frontline workers.

The timing matters. Brazil's Lei Geral de Proteção de Dados, in force since September 2020, imposes strict requirements on how public bodies retain and delete personal data, including images captured in public spaces. Holding duplicate files containing identifiable faces or vehicle plates is not a neutral IT inefficiency — it is a potential compliance liability. With the federal government in Brasília pressing municipalities to demonstrate LGPD adherence before the end of 2026, the window for São Paulo to act is narrowing fast.

Where the Problem Lives — and Who Owns It

The duplication burden is concentrated in two city systems. The Centro de Operações São Paulo, the real-time urban monitoring hub based on Rua Líbero Badaró in the historic centre, ingests imagery from more than 4,000 camera feeds across the city. Separately, the Secretaria Municipal de Habitação uses photographic records to document progress on housing interventions in communities including Heliópolis, in the Ipiranga district, and Paraisópolis, in Morumbi. Both operations generate large volumes of images daily, and neither has a unified deduplication protocol.

The Instituto de Tecnologia e Sociedade do Rio de Janeiro, which monitors digital governance across Brazilian cities, has noted that municipal image archives in major urban centres tend to grow at roughly 40 percent annually once smart-city camera networks reach a certain density — a pace that most deduplication workflows, built for smaller datasets, cannot match. São Paulo's camera network crossed the 4,000-node threshold in early 2025.

Private sector players are circling. At least three São Paulo-based tech firms with roots in the Cubo Itaú startup hub in Faria Lima have pitched deduplication-as-a-service contracts to the city. The procurement decisions that follow will determine not only which technology the city adopts but who holds administrative access to sensitive urban imagery for years to come — a question that civil liberties groups on Avenida Paulista have already flagged in public consultations this year.

The Decisions That Cannot Wait

Three choices are immediately in front of City Hall. First, Mayor Ricardo Nunes' administration must decide whether to handle deduplication in-house, using the existing technical team at the Secretaria de Inovação, or to contract externally. An in-house build preserves data sovereignty but requires hiring — the secretariat currently lists seven open positions for data engineers on the Portal do Servidor, São Paulo's public-sector jobs board. An external contract moves faster but introduces the governance questions around access and audit trails.

Second, city officials need to establish a retention schedule before any deletion campaign begins. Under LGPD Article 16, personal data can only be held as long as strictly necessary. Agreeing on what counts as necessary — one week of redundant surveillance stills? One month? — requires a legal opinion from the Procuradoria-Geral do Município and sign-off from the city's data protection officer, a post that has been occupied on an interim basis since March 2026.

Third, the city must decide whether to publish its deduplication methodology openly. Transparency advocates argue that releasing the technical criteria — what hash functions are used, what perceptual similarity thresholds trigger deletion — would allow independent auditors to verify that no non-duplicate images are accidentally removed. Several European cities, including Amsterdam and Lisbon, have published comparable documentation. São Paulo has not committed to doing so.

The practical timetable is tight. Federal LGPD audit cycles for municipalities are expected to begin in the fourth quarter of 2026. That gives the Nunes administration roughly three months to move from acknowledged problem to documented policy. The decisions made in July and August will determine whether São Paulo enters that audit window with a credible governance framework — or a storage bill and a legal exposure it cannot easily explain.