news
Johannesburg's Digital Records System Collapses Under Duplicate Files Crisis
Years of fragmented city systems, load shedding-driven data corruption, and a migration crisis that overwhelmed municipal databases have left Joburg's image archives riddled with duplicate files nobody can easily trace or remove.
How we reported this
The City of Johannesburg's digital infrastructure has a problem that predates the current ANC-DA coalition government by more than a decade: its image and document repositories are bloated with millions of duplicate files, degraded scans, and mismatched records that slow down everything from rates billing to housing applications. Officials at the Pikitie Street civic centre in Braamfontein have been quietly grappling with the fallout for years. Now, with a formal deduplication programme reportedly under discussion inside the city's Group ICT directorate, residents and civil-society organisations are asking how things got this bad in the first place.
Understanding that question matters more than ever in 2026. The coalition government that paired the ANC and the DA in Gauteng after the 2024 elections made modernising city services a centrepiece of its shared platform. Cleaning up duplicated digital records, ranging from property photographs in the City's cadastral system to identity documents held by the Johannesburg Social Development Department, is a precondition for almost every other reform on that list. You cannot automate rates notices or digitise housing queues if the underlying image data is corrupted or duplicated three times over.
A Crisis Years in the Making
The roots of the problem stretch back to at least 2008, when the City of Johannesburg began pushing departments to scan and digitise paper records without providing a unified storage standard. Different directorates, Water and Sanitation, Housing, the Johannesburg Roads Agency, purchased separate server infrastructure and used incompatible file-naming conventions. When records were later consolidated onto shared drives managed from the Thuso House complex in Braamfontein, the result was predictable: thousands of image files appeared multiple times under different names, with no automated tool in place to detect or merge them.
Load shedding made everything worse. During the worst periods of Stage 6 cuts in 2022 and 2023, interrupted write cycles corrupted image batches mid-upload, forcing IT staff to re-upload entire folders. Those re-uploads frequently created secondary copies sitting alongside the originals, untagged and untracked. The Johannesburg Development Agency, which manages urban renewal projects across areas including the Maboneng Precinct on Fox Street and the Inner City Improvement District around Diagonal Street, has told stakeholders in past annual reports that its project-photo archives alone contain redundant files running into the tens of thousands.
Migration pressures compounded the storage sprawl. Between 2019 and 2024, the City processed a sharp rise in applications for informal settlement upgrades and social-housing allocations tied to new arrivals from Zimbabwe and Mozambique, particularly in densely populated areas like Hillbrow and the southern sections of Soweto. Each application required scanned supporting documents, identity documents, proof of address, utility photographs. Many applicants submitted multiple applications across different offices, and the City had no real-time deduplication layer to catch the resulting redundant image sets before they were written to disk.
What a Fix Actually Looks Like
Deduplication is not a single button. The technical approach involves hashing algorithms that generate a unique fingerprint for each image file and then flag identical fingerprints for review, a process that, in a repository the size of the City's, can take weeks of compute time and requires staff trained to adjudicate edge cases where two images are near-identical but not exact copies. Municipal IT procurement in South Africa requires Treasury-compliant tender processes under the Municipal Finance Management Act, meaning even a software licence for a commercial deduplication tool must go through a formal bid adjudication committee before it can be deployed.
The City's Integrated Development Plan for 2025-2030, a public document, flags digital governance as a cross-cutting priority, though it stops short of committing budget figures to any specific data-cleaning exercise. Civil-society groups monitoring the process, including the Centre for Public Accountability, which operates out of offices on Jan Smuts Avenue in Parktown, have pushed for a published timeline and an independent audit of the repositories before any deduplication work begins, to establish a baseline and prevent accidental deletion of legally significant records.
For ordinary Joburgers, the practical upshot is straightforward: if you have submitted scanned documents to a City of Johannesburg office more than once, for a water-account query, a rates objection, or a housing application, hold onto your originals. Until the deduplication programme is complete and audited, there is no guarantee that the City's systems hold a clean, retrievable copy of what you gave them.