Building a Data Warehouse for a Million Learners: Our Partnership with Imagine Worldwide

Some partnerships shape who you become as a company. For us, Imagine Worldwide is one of them.

We started working with Imagine back in 2022, when our team was still part of Ona. Four years later, the relationship has grown alongside the organisation — and the data behind it.

Who is Imagine Worldwide

Imagine Worldwide is a non-profit on a mission to solve the foundational literacy and numeracy crisis for children across Sub-Saharan Africa. Their approach is elegant in concept and ambitious in scale: put a solar-powered tablet, loaded with adaptive learning software in the child's national language, into the hands of every learner in a primary school. Each child works independently for about 150 minutes a week on reading and maths, with software that adapts to where they are in their learning journey.

The model is working. Imagine partners directly with the governments of Malawi, Tanzania, and Sierra Leone, and is currently deployed in roughly 1,800 schools, reaching over a million learners and scaling rapidly. In Malawi, they are on a path to support every primary school in the country — close to 3.8 million children. In 2025 they were awarded the inaugural Global EdTech Prize and have been named into the Audacious Project cohort — recognition that the model is not just promising but proven.

It is a clear values match. Imagine is doing serious work to expand access to quality education at a scale that actually matters, and we want to be useful to organisations like that.

The data challenge

Operating at this scale produces a particular kind of data problem. Every tablet is, in effect, an IoT device — generating fine-grained interaction data each session: which lessons a child completed, where they struggled, how long they spent on a task, what the device itself was doing. Multiply that by hundreds of thousands of learners across thousands of schools across multiple countries, and you get volumes and frequencies that traditional analytics warehouses start to strain under.

But the operational picture is more than just tablet telemetry. To answer the questions Imagine actually cares about — Are children learning? Where are deployments succeeding? Where do we need to intervene? Are we on track against country-level plans? — you need to combine tablet data with school operations data, with country programme plans, with device-fleet management, with internal targets. Different shapes, different cadences, different owners.

What we built

Since 2022 we have designed and continuously evolved the data architecture behind this. Through 2025 and into 2026, a significant chunk of the work has been a careful migration of the warehouse to ClickHouse, which is now the core data warehouse for the organisation. It comfortably handles the terabytes of tablet and operational data, with query performance that lets the team explore questions interactively rather than waiting on overnight runs.

On top of ClickHouse sits a dbt-based analytics engineering codebase that does the heavy lifting: cleaning, modelling, and transforming raw events into the metrics that programme teams, country leadership, and funders rely on. And we deliver those insights through Akuko, where dashboards bring the data within reach of the people making decisions, not just the analysts.

Stabilising a migration of this size, while keeping the warehouse fully operational, has been one of the more rigorous engagements we have run — and probably the one that has taught us the most about managing data at real scale and in support of a truly data-driven organisation.

Looking forward

We are continuing to work alongside Imagine Worldwide on what comes next. The roadmap is genuinely exciting:

• More ambitious insight generation — pushing beyond reporting into the questions that change programme decisions.
• New geographies and new integrations — including pulling ERP and finance data into the warehouse, so operational and educational data live in the same picture.
• AI agents for self-service insight and automation — making the warehouse useful to non-technical users and reducing manual analytical lift across the team.

We are proud of this partnership, and proud of what Imagine has built. We are looking forward to seeing how they take this data forward — and how many more children get access to quality education on the African continent as a result.

"Ona Insights brings deep institutional knowledge of our data systems and the commitment to go the extra mile when it counts — truly a reliable partner for any organisation working to deliver impact at scale in challenging environments."

— Johan Wannenburg, Director of Analytics at Imagine Worldwide