Nairobi's Digital Archive Problem: How the City Stacks Up Against Lagos, Accra and Nairobi's Own Ambitions on Duplicate Image Data

Nairobi's city government and its ecosystem of public-sector technology vendors are sitting on a sprawling, largely unaudited mess of duplicate digital images — property survey photographs, ID registration scans, informal settlement mapping files — and the bill for storing them is quietly eating into already strained budgets under the county's IMF-constrained fiscal envelope.

The problem crystallised in the first half of 2026 when the Nairobi City County ICT directorate began an internal audit of its digital asset management systems ahead of a planned migration to a new cloud infrastructure platform. The audit, according to documents circulating among county technology vendors in Westlands, found that duplicate and near-duplicate image files were consuming a disproportionate share of the county's contracted cloud storage. Exact figures from that audit have not been made public, but the review prompted an urgent procurement notice posted to the county's supplier portal in May 2026.

The timing matters. Kenya's national government is under pressure from the International Monetary Fund to reduce recurrent expenditure, and county governments have felt that squeeze directly. Paying cloud storage fees for redundant data is precisely the kind of avoidable cost that fiscal reformers — and the Gen Z-era tax accountability movement that shook Nairobi's streets in 2024 — have made politically toxic.

What Lagos and Accra Got Right Earlier

Lagos State in Nigeria moved to address duplicate image data in its land registry digitisation programme starting in 2023, contracting local firm CRC Credit Bureau and a consortium of Lagos-based software developers to implement perceptual hashing tools — algorithms that detect visually similar images even when file names or metadata differ. By early 2025, Lagos State's Bureau of Lands and Survey reported it had cut its scanned-document storage footprint by consolidating redundant images across its property title database. Accra's Lands Commission in Ghana similarly adopted automated deduplication as part of a World Bank-supported land administration modernisation project, with implementation beginning in 2024.

Nairobi is roughly 18 to 24 months behind those two cities on this specific technical challenge, according to technology consultants who work across all three markets. The gap is not primarily a skills problem — iHub in Kilimani and the Chandaria Business Innovation and Incubation Centre at the University of Nairobi both have resident developers who build image-recognition pipelines commercially. The gap is procurement and political will.

The informal settlement upgrading programme run under the Kenya Informal Settlements Improvement Projects, which operates in areas including Mukuru kwa Njenga and Mathare, generates thousands of aerial survey and ground-level photographs per enumeration cycle. Community surveyors working with mobile data collection tools frequently upload the same image multiple times due to poor connectivity, and the deduplication layer in the current data pipeline is manual and intermittent. A sector coordinator familiar with the programme's data workflows described the situation in general terms at a GIS Kenya meetup held in April 2026 at the iHub offices on Senteu Plaza, Kilimani — though no specific figures were shared publicly at that event.

The Cost Calculation and What Comes Next

Commercial cloud storage pricing gives a rough sense of the stakes. Amazon Web Services S3 standard storage costs approximately $0.023 per gigabyte per month as of mid-2026. Organisations storing even a few tens of terabytes of image data — well within the scale of a city government's property and population registration systems — can accumulate storage bills running into tens of thousands of dollars annually before deduplication is applied. For Nairobi County, converting those dollar costs to Kenya shillings at the prevailing rate above Ksh 130 to the dollar makes the arithmetic uncomfortable.

The practical path forward involves three steps that peer cities have already mapped: deploying perceptual hashing or MD5 checksum deduplication at the point of data ingestion rather than retrospectively; mandating that any new county digital system procurement include deduplication as a baseline technical requirement; and running a one-time retrospective clean-up of existing archives. The Nairobi Metropolitan Services, which oversees several digitisation initiatives in the greater metro area including the commuter rail data systems along the Syokimau corridor, is positioned to pilot this approach at scale if the county and national government can agree on a shared data governance framework — conversations that, as of early July 2026, are ongoing but unresolved.