São Paulo's Digital Archives Plagued by Thousands of Corrupted Ghost Images

Thousands of duplicate and broken image files have accumulated inside São Paulo's municipal digital systems, slowing public databases, inflating storage costs, and creating headaches for the civil servants who manage them. The problem is not new, but pressure to resolve it has sharpened in recent months as the city's transparency portal, accessible through the Prefeitura de São Paulo's official website, struggles under the weight of redundant visual assets attached to urban planning documents, public works filings, and social program records.

The issue surfaced publicly during a working session of the Câmara Municipal de São Paulo in June 2026, when councillors from the infrastructure committee raised concerns about the performance of the GeoSampa mapping platform, which the city uses to manage spatial data across all 96 administrative subprefeituras. Technical staff told the committee that a significant share of the platform's storage load was tied to image duplication, photographs of street conditions, drainage infrastructure, and building permits that had been uploaded multiple times without a deduplication protocol in place.

Why This Matters Beyond Slow Load Times

Storage is not cheap. Municipal cloud infrastructure contracts in São Paulo, processed through the Secretaria Municipal de Inovação e Tecnologia, run into the tens of millions of reais annually. Technology specialists outside government say that image redundancy in large urban databases typically inflates storage demand by 15 to 30 percent, though the exact figure for São Paulo's systems has not been officially published. The practical consequence is that engineers and archivists spend hours manually identifying and removing duplicate files rather than building new tools.

The flooding documentation system is one area where the duplication problem has real-world stakes. After each major rain event, and São Paulo recorded more than 20 flood-related emergencies in the first half of 2026 alone, field inspectors upload photographs of damaged streets and drainage blockages to the Defesa Civil São Paulo database. Without automated image-matching tools, the same photograph can exist in three or four separate records simultaneously, muddying the damage assessment reports that feed budget requests to the state and federal governments.

Specialists in digital public administration at the Fundação Getulio Vargas campus on Rua Itapeva, in Bela Vista, have been tracking the issue as part of broader research into municipal data governance across Latin American cities. Their position, stated in a working paper circulated earlier this year, is that the absence of a standardised duplicate-detection protocol puts São Paulo behind comparable major urban centres in Europe and North America. The research does not assign blame to any individual office but argues the gap is structural.

What City Hall Is Saying, and What Comes Next

The Secretaria Municipal de Inovação e Tecnologia has acknowledged the problem in internal communications reviewed by this newspaper, though no formal public statement has been issued. Officials within the secretariat have indicated that a procurement process for image deduplication and asset management software is being prepared, with a target to publish the edital, the public tender notice, before the end of the third quarter of 2026. If that timeline holds, a contract could be awarded by December.

The Instituto de Pesquisas Tecnológicas, the state-linked research body headquartered in Cidade Universitária in the west of the city, has been named informally as a potential technical partner for any pilot programme. Researchers there have experience with large-scale image classification systems, including work done in partnership with SABESP on infrastructure monitoring.

For ordinary Paulistanos, the immediate ask is modest: if you submit documents or images to any city portal, whether for a building permit in Pinheiros, a noise complaint in Mooca, or a social services application in Cidade Tiradentes, avoid uploading the same file multiple times. The digital backlog those duplicate uploads create has a cost that ultimately lands on the municipal budget. City technologists say a public-facing guidance update is being drafted for the SP156 service platform, São Paulo's main citizen request channel, and should appear before August.