Continental Research Architecture for Data Linkage in Education
Africa federates health data at continental scale. The Africa CDC Central Data Repository, launched in January 2026, aggregates surveillance, laboratory, and programme data from national health systems across the continent. The accompanying Continental Health Data Governance Framework provides the institutional and legal template for sovereignty-preserving cross-border data sharing.
Education has no comparable infrastructure. Country-level education data remains siloed in incompatible national Education Management Information Systems (EMISs), invisible to cross-jurisdictional research, inaccessible to continental policy, and unusable for the evidence-based financing that Africa's education systems urgently require.
Africa's Digital Public Infrastructure for Education (DPI-Ed) changes this landscape. The RESPECT Platform, as the first reference implementation of Africa's DPI-Ed, generates standardized, curriculum-aligned learning data by default — continuous evidence of what learners actually learn, produced through standardized interfaces across every participating country. For the first time, comparable education data will exist across African jurisdictions. The question is whether that data will fragment along national boundaries — as legacy EMIS data has — or be federable from the outset.
CRADLE — Continental Research Architecture for Data Linkage in Education — is a 24-month, $10M ± $2M research program that will design, prototype, and validate the architecture, governance framework, and operational policies for the back-end of Africa's DPI-Ed across participating Member States, as a complement to RESPECT as that DPI-Ed's user-facing front-end
CRADLE will deliver a working Malabo Convention-compliant prototype federating DPI-Ed-generated data across six pilot countries, a validated continent-scalable architecture specification, a data governance framework addressing anonymization, aggregation, sovereignty, and tiered access control, and peer-reviewed research.
Africa's multi-sovereign environment — 50+ countries, each with its own data authority — combined with the Malabo Convention's prohibition on cross-border Personally-Identifying Information (PII) flows and the offline-first, low-bandwidth conditions of Africa's schools, generates six architectural constraints that shape every design decision: data sovereignty, governance separation, right-sized complexity, open data formats, EMIS compatibility, and offline-first low-bandwidth operation (Section 4 derives each from its structural premises). Three governance purposes are currently identified — operational, research, and RBF4Ed (Results-Based Financing for Education) — with the architecture designed to accommodate additional pipelines as new purposes emerge.
The federated data architecture CRADLE produces will serve as the continental intelligence layer for the entire Breakthrough System. Researchers will gain cross-jurisdictional datasets revealing patterns invisible within any single country. App developers will identify performance variations across deployment contexts. School Leaders will benchmark against peers. RBF4Ed's evidence will become exponentially more valuable when comparable across jurisdictions. The PREMIER Institute's research projects will draw on CRADLE's federated data for training and validation at continental scale. And every additional country that joins the federated-data network will make the whole more valuable — a network effect that strengthens the case for V&P_Core adoption.
Africa's education systems generate data. Every country maintains some form of Education Management Information System (EMIS) — ranging from paper-based registers to partially digitized platforms — recording enrollment figures, examination results, teacher attendance, school infrastructure inventories, and other administrative indicators. These systems were designed for national administrative purposes: reporting to Ministries, satisfying donor requirements, and tracking inputs (teachers hired, textbooks distributed, schools built).
What these systems produce is a record of what was provided. What they do not produce — and cannot currently support — is cross-jurisdictional evidence of what was learned. Which teaching approaches produce measurably better outcomes across linguistically similar regions? How do learning trajectories differ between countries that adopted the same curriculum framework? What can a country struggling with numeracy learn from a neighboring country with similar demographics that achieved stronger results? These questions require comparable, curriculum-aligned outcome data across multiple jurisdictions. That data does not exist in any integrated form.
The fragmentation is structural. Africa's EMISs differ in data models, collection methods, indicator definitions, reporting frequencies, and technical platforms. Some countries use DHIS2 (originally a health platform, increasingly adapted for education). Others use proprietary systems built by different vendors under different donor-funded projects. Some still rely on paper-based data collection with periodic digitization. There is no shared data standard for education across the continent, no common indicator framework, and no institutional mechanism for cross-border data comparison.
This is the same structural barrier that health data faced before DHIS2 achieved continental adoption — and that health is still working to resolve at the federation level. Education is a generation behind health in data infrastructure maturity.
Africa's health data infrastructure provides the closest precedent for what CRADLE will build. Three components of health's experience are directly instructive.
DHIS2 is the world's largest health management information system, deployed in over 100 countries, with particularly deep penetration across Africa. Its organizational hierarchy model — representing administrative structures from national level down to individual health facilities — enables aggregation at any level. DHIS2's success demonstrated that a shared, open-source data platform can achieve continental adoption despite heterogeneous national systems, provided the platform accommodates local variation within a common framework.
Africa CDC's Central Data Repository (CDR), launched in January 2026 after 15 months of feasibility study and prototype validation, federates surveillance, laboratory, and programme data from national health systems across the continent. The CDR's foundational design principle is federation: countries contribute data to a continental view while retaining full ownership of and access to their own data. The CDR demonstrates that continent-scale data federation is technically achievable and politically viable.
The Continental Health Data Governance Framework, set for AU endorsement alongside the CDR, provides the institutional template for cross-border data sharing: data-sharing agreements, sovereignty protections, access controls, and Malabo Convention compliance. The Framework addresses the political and legal challenges of multi-jurisdictional data governance — challenges that are as formidable as the technical ones.
Health's experience also reveals what to address from the outset. Health data federation has been complicated by inconsistent facility-level reporting, difficulty distinguishing zero cases from missing data, and manual data entry errors — problems rooted in the fact that health data originates through manual collection and transcription. Health also achieved continental DHIS2 adoption before developing shared indicator definitions, creating comparison problems that are still being resolved. And the Continental Health Data Governance Framework is being developed decades after DHIS2 deployment began — a sequencing problem that CRADLE can avoid.
CRADLE's strategy is to learn from health's architectural patterns and governance models while addressing these known problems from the outset.
Africa's DPI-Ed fundamentally changes the education data landscape. Legacy EMIS systems are administrative databases — they record what was provided (inputs). Africa's DPI-Ed generates continuous, curriculum-aligned evidence of what was learned (outcomes). Every interaction between a learner and a RESPECT Compatible App produces standardized learning data through standardized technical interfaces.
This distinction is consequential for federation. DPI-Ed-generated data is born digital, standardized at the point of creation, and curriculum-aligned by design. It eliminates the manual collection and transcription errors that degrade health data quality. It produces comparable data structures across every country deploying the RESPECT Platform. And it generates continuous evidence — ongoing learning signals, measured against curriculum standards — where legacy EMIS data provides only periodic snapshots of administrative inputs. The data is standardized and curriculum-aligned from the moment of creation on the learner's device; its arrival at the country-level data store, however, is asynchronous — governed by the intermittent connectivity and bandwidth constraints that characterize Africa's low-resource school environments (see Section 4.6). Any federation architecture must be designed for data that arrives in delayed batches, not in real-time streams.
For the first time, comparable education outcome data will exist across the countries deploying Africa's DPI-Ed. CRADLE's purpose is to ensure that this data is federable from the outset — that the architecture supports continent-scale analysis while preserving the national sovereignty that makes participation possible.
Five developments have converged to make CRADLE feasible and urgent: Africa CDC has proven continent-scale data federation; V&P_Core's six pilot countries will generate standardized learning data within CRADLE's programme timeline; the Malabo Convention has entered into force (June 8, 2023), providing the continental legal framework for data protection and cross-border data flows; RBF4Ed depends on cross-jurisdictional evidence that CRADLE's federation layer provides; and the AU Assembly's adoption of Decision Assembly/AU/Dec.973(XXXIX) creates the institutional mandate for CRADLE's continental data architecture.
Three facts about Africa's education landscape shape every architectural decision CRADLE makes.
First, Africa's education system spans 50+ sovereign nations. Each country maintains its own Education Management Information System, its own curriculum standards, its own data protection authority, and its own Ministry of Education with sovereign jurisdiction over its learners' data. Any continental education data architecture must serve all of these countries — and must earn and sustain each country's voluntary participation.
Second, the Malabo Convention — the African Union Convention on Cyber Security and Personal Data Protection, which entered into force on June 8, 2023 — prohibits raw Personally Identifiable Information (PII) from crossing national borders without explicit consent. Within a single country, a Ministry of Education may use PII-rich learner data for any lawful purpose — dashboards, interventions, reporting. The moment that data is to serve a purpose beyond the source country, the Malabo Convention constrains what may leave and in what form.
Third, the schools where Africa's DPI-Ed operates are predominantly low-resource environments — characterized by intermittent or absent connectivity, expensive and scarce bandwidth, low-end shared devices, and unreliable power. The DPI-Ed front-end must function fully offline, synchronize opportunistically when connectivity appears, and minimize the data volume transmitted over expensive mobile networks. Learning data is generated on devices that may remain disconnected from the country-level back-end for days or weeks at a time. Any continental data architecture must be designed for a front-end that operates under these conditions — receiving data in delayed, compressed batches rather than in real-time streams.
Together, these three facts generate six architectural constraints that any continental education data system must satisfy. These are not design preferences — they are structural consequences of building for 50+ sovereign nations under continental data protection law in low-resource deployment environments.
Right-Sized Complexity. Africa's 50+ AU member states span three orders of magnitude in school-age population — from the Seychelles (under 20,000 students) to Nigeria (over 40 million). The architecture must scale down as gracefully as it scales up, serving the smallest countries with operationally simple and cost-effective infrastructure while accommodating the largest countries' data volumes — all within a single architectural framework.
Open Data Formats. Federation across 50+ sovereign systems requires format interoperability that proprietary solutions are unable to guarantee. Open, internationally-standardized data standards ensure that any country's data can participate in the federation regardless of its implementation choices. For DPI-Ed-generated learning data, IEEE's xAPI provides the open standard for learning event capture.
The remainder of this section addresses each constraint and its architectural consequences.
CRADLE's foundational design principle is federation of sovereign data. Each School Leader — whether a Ministry of Education or the leader of a private, NGO-based, or faith-based school system — retains sovereignty over its own data. CRADLE defines the interfaces, protocols, and governance through which sovereign data holders choose to share specified data streams, at specified levels of detail, under specified conditions.
This is the same architectural principle that governs DHIS2's most successful deployments and Africa CDC's Central Data Repository.
Africa's DPI-Ed currently serves three distinct governance purposes, each requiring its own pipeline. These are the design-driving set — the pipelines CRADLE's architecture must support from the outset — but the architecture must accommodate additional pipelines as new purposes emerge.
Operational Pipeline. Ministry-facing dashboards, continental reporting, and automated intervention triggers. Within each country, this pipeline operates on PII-rich data — individual learner progress, assessment scores, attendance patterns — enabling Ministries to act on what they see. At the continental level, this pipeline federates pre-computed statistical summaries: completion rates by district, assessment score distributions by subject, engagement trends. Raw PII stays within the country; the continental view is assembled from aggregates. The Operational Pipeline is always active for every participating country.
Research Pipeline. Cross-jurisdictional research on anonymized, statement-level data, governed by CRADLE's consent framework. Operational aggregates answer questions that someone has already thought to ask — they are pre-computed summaries at defined levels of detail. Research requires access to finer-grained data to discover patterns that were not anticipated: time-on-task correlations with assessment outcomes across francophone West Africa, or the interaction between class size and app effectiveness across countries with similar curricula. The Research Pipeline exists for this purpose. Countries opt in to sharing anonymized statement-level data for specific approved studies, under time-bounded access with audit trails. PII is anonymized before it leaves the source country. This is CRADLE's primary pipeline — the one that produces the cross-jurisdictional education intelligence that has never existed for Africa.
RBF4Ed Pipeline. Auditable outcome evidence for Results-Based Financing for Education (RBF4Ed), governed by the GEOS Organization and certified by GEOSors (see Essay 7). This data triggers financial disbursement — it must be tamper-evident, independently auditable, and resistant to gaming. The RBF4Ed Pipeline operates at the country level and is expected to terminate there: finance-grade evidence is not federated continentally. The GEOS Organization defines the integrity standards; RBF4Ed converts certified evidence into funding flows.
CRADLE's architecture must ensure clean separation among pipelines — so that operational aggregation does not compromise the fine-grained data access that research requires, research access does not compromise finance-grade integrity, and finance-grade audit requirements do not constrain the research and policy value of the educational data stream. CRADLE is directly responsible for the Research Pipeline and provides the governance framework that the Operational Pipeline's continental aggregation operates within. The RBF4Ed Pipeline is architecturally independent — designed and governed by the GEOS Organization under RBF4Ed — but CRADLE's country-level data architecture must ensure that the data store exposes the interfaces each pipeline requires.
A federation architecture that requires each participating country to operate a complex data infrastructure imposes costs that are fixed regardless of the number of students served. For the smallest countries, the operational overhead of a data federation node — servers, security, database administration, governance compliance — could consume resources disproportionate to the data volume it processes. The per-country federation node must therefore be operationally simple enough for a small country's team to run and cost-effective enough for a small country's budget to sustain, while accommodating large countries whose data volumes require more capable infrastructure — all within a single architectural framework. This is the same right-sizing constraint that governs RESPECT's deployment model (see Essay 3, Section 5) and that the IMPACT Board's Impletor and DiPian certification pipelines address for human capacity: the infrastructure must be operable by the professionals that the system trains.
For DPI-Ed-generated learning data, the xAPI (Experience API) standard provides the open specification for capturing, storing, and federating learning events. Every RESPECT Compatible App produces xAPI-formatted learning statements through V&P_Core's standardized interfaces — meaning that learning data is born federable, in a format governed by an open community standard. For EMIS-originated administrative data, open API specifications — developed through the BEINGS Planet-Project's EMIS Interoperability building block — provide the integration standard. BEINGS is expected to produce spec-wrappers for both DPI-Ed and EMIS interfaces, enabling CRADLE's federation layer to ingest data from both streams through a common, openly specified interface.
The heterogeneity of Africa's 50+ national EMIS systems defines the integration challenge. The BEINGS Planet-Project's EMIS Interoperability building block specification is designed to address this "post-entrenchment landscape" (see also Essay 3, Section 7).
Africa's DPI-Ed generates learning data on devices in schools where connectivity is the exception, not the norm. The front-end must treat offline as the default operating state — recording learning events, managing rosters, delivering content, and running local analytics from local data — and synchronize with the country-level back-end when connectivity becomes available. Content must be distributable through local peer-to-peer transfer between devices on the same network, so that a lesson downloaded once by one device can reach every device in the school without additional external bandwidth consumption. Data uploads to the back-end must be compressed and batched, minimizing the volume transmitted over expensive mobile connections.
This constraint has three consequences for CRADLE's architecture:
Asynchronous, bursty data ingestion. The country-level data store cannot assume a steady stream of learning events. Devices may synchronize after hours, days, or weeks offline, producing large batch uploads — potentially from thousands of schools reconnecting simultaneously after a connectivity window. The ingestion layer must handle delayed, high-volume batch arrivals without data loss or performance degradation.
Data completeness as a first-order architectural concern. If devices are intermittently connected, the country-level data store may be temporarily — or permanently — incomplete. Some learning events may never arrive (device lost, battery dead, student transfers). The architecture must distinguish between "no learning events occurred" and "learning events occurred but have not yet been received" — the same zero-vs-missing problem that health data federation has struggled with (Section 2), intensified by the offline-first operating model. Every aggregation, dashboard, and federation summary must carry completeness indicators: what fraction of expected devices have synchronized, and how recently.
Integrity under delayed delivery. The RBF4Ed Pipeline (Section 4.2) requires tamper-evident, auditable outcome evidence. When that evidence arrives in delayed batches from devices that were offline, the chain of custody is harder to verify than for data transmitted in real time. The architecture must provide cryptographic assurance that learning events were recorded at the time of the interaction and have not been altered between recording and receipt — even when days or weeks separate the two.
The six constraints derived above — data sovereignty, governance separation, right-sized complexity, open data formats, EMIS compatibility, and offline-first low-bandwidth operation — are requirements, not aspirations. Any architecture that fails to satisfy all six is non-viable for continental-scale education data federation across Africa. This section describes one candidate architecture that satisfies all six, offered as a starting hypothesis for CRADLE's research programme to test, refine, or replace. The purpose is to demonstrate that a viable architecture exists — and to give the research programme a concrete design to evaluate against alternatives.
Governance separation through per-purpose read paths. All pipelines read from the same country-level data store but are structurally separated. The Operational Pipeline computes aggregations locally and federates summaries to the continental hub. The Research Pipeline anonymizes statement-level data within the country and provides time-bounded, consent-governed access under CRADLE's research framework. The RBF4Ed Pipeline extracts tamper-evident outcome evidence for certification by the GEOS Organization. Each pipeline has its own access controls, its own audit trail, and its own governing authority. The architecture supports additional pipelines by exposing the country-level data store through governed, purpose-specific interfaces.
Right-sized complexity through a modular monolith. The country-level back-end is built as a single deployable application with strict internal module boundaries — a learning-events module (xAPI ingestion and validation), an analytics module (aggregation and query), a decision-support module (dashboards and alert rules), a communications module (multi-channel message delivery for automated interventions), and a credentials module (W3C Verifiable Credential issuance and verification). A single trained implementor can deploy and operate a country-level instance. For the largest countries, specific modules (the ingestion tier) are extracted as separately scaled services when load demands it — the same architecture, different deployment topology.
Open data formats throughout the stack. xAPI (IEEE 9274.1.1) is the data format from learner device to country-level LRS to continental federation hub. Every learning interaction produces xAPI-formatted statements through standardized interfaces; no proprietary translation layer is required. The analytical data store ingests xAPI natively. Dashboards query xAPI-derived aggregations. The federation protocol transmits xAPI-compatible summaries. For EMIS-originated data, open API specifications — developed through the BEINGS Planet-Project — provide the integration standard, with country-specific adapters translating each national EMIS into a common, openly specified interface.
EMIS compatibility through GovStack-compatible interfaces. The EMIS interface is a standardized specification that can be implemented for all participating countries. Country-specific adapters accommodate EMIS heterogeneity; the interface makes it easy to ensure federation-level comparability. BEINGS' spec-wrappers provide the adapter specifications for both DPI-Ed and EMIS data streams, giving the federation layer a common integration surface regardless of each country's existing systems.
Offline-first resilience through asynchronous batch ingestion. The DPI-Ed front-end operates offline by default: learning events are recorded on-device at the moment of the interaction, stored locally, and transmitted to the country-level back-end in compressed batches when connectivity becomes available. Content reaches devices through local peer-to-peer transfer between nearby devices, consuming external bandwidth only once per school for each content item. The country-level ingestion layer accepts delayed, high-volume batch uploads — potentially from thousands of schools reconnecting simultaneously — without assuming real-time connectivity or server-dependent front-end workflows. Every aggregation and federation summary carries completeness metadata: the fraction of expected data sources that have synchronized and the recency of each source's latest sync. The RBF4Ed Pipeline's tamper-evidence requirements are satisfied by cryptographic signatures applied on-device at the time of recording, verifiable regardless of transmission delay.
Every component in this architecture already exists as production-grade Free and Open Source Software (FOSS), proven at scales that exceed Africa's requirements. The xAPI Learning Record Store is Ralph (Python/FastAPI, AGPL-3.0), already deployed in Open edX Aspects serving millions of learners worldwide. The analytical data store is ClickHouse (Apache 2.0), a columnar database processing trillions of rows in production at Yandex, Cloudflare, and hundreds of other organizations. Decision-support dashboards are provided by Apache Superset (Apache 2.0), the business intelligence platform used by Airbnb, Lyft, and Dropbox. The communications module uses Chatwoot (MIT) for multi-channel message routing and Africa's Talking — a Nairobi-based company — for SMS delivery across African mobile network operators. Verifiable credentials are handled by Walt.id SSI Kit (Apache 2.0), implementing the W3C Verifiable Credentials standard. The proposed architecture composes these proven components — each individually battle-tested at scales exceeding the projected education workload — into a federated system that satisfies all six constraints by construction.
The proposed architecture's internal module interfaces are expected to reuse the high-level interfaces/APIs of India's DPI-Ed program, notably Sunbird — adapting field-tested interfaces to an architecture shaped by Africa's multi-sovereign constraints.
The candidate architecture above is the starting hypothesis. CRADLE's research programme — described in the next section — will evaluate this design against alternatives, stress-test its assumptions, and produce the validated architecture specification that the continent will build on.
CRADLE is a 24-month, $10M ± $2M research programme in two phases, with a go/no-go gate at Month 12. Phase 1 (Months 1–12, $6M ± $1M) designs the federated architecture, develops a working Malabo-compliant prototype across at least two pilot countries, and establishes the governance framework. Phase 2 (Months 13–24, $4M ± $1M) validates the architecture across all six pilot countries, stress-tests the governance framework, and prepares the scaling pathway to 50+ countries. The programme's phasing, milestones, dependencies, and risk mitigations are detailed in the CRADLE Database Project Plan.
Participation in CRADLE requires each pilot country to provide a Ministry-designated data liaison, a data-sharing agreement governing federation terms, and operational access to the country's EMIS for integration assessment. Phase 1's initial prototype countries must have data-sharing agreements and EMIS access in place within the programme's first six months; remaining pilot countries must have them in place by Phase 2 entry (Month 13). CRADLE's engineering team provides the federation node infrastructure, adapter development, and technical training; the country provides the institutional authorization and local coordination that make deployment possible.
CRADLE's federated architecture provides the continental intelligence layer on which multiple Breakthrough System components depend.
Cross-jurisdictional research at continental scale. Researchers will access federated education data revealing patterns invisible within any single country. The data is born digital, standardized, and curriculum-aligned — enabling analyses at a quality and scale that have never been possible for African education.
Evidence amplification for RBF4Ed. Cross-jurisdictional outcome evidence is exponentially more valuable for results-based financing than single-country data. CRADLE enables comparative benchmarking that strengthens the evidentiary case for RBF4Ed disbursements — measuring outcomes across jurisdictions, identifying what works where, and providing the continental intelligence that independent finance facilities require.
Training data for the PREMIER Institute. Four of five PREMIER Institute research projects — Easy Personalized Learning, Easy Knowledge Assessment, Easy Text Localization, and Easy Courseware Gamification — are expected to draw on CRADLE's federated data for model training and validation. CRADLE provides the continent-scale, sovereignty-preserving data access that makes PREMIER's research tractable (see the PREMIER Institute Project Plan).
Sovereign data infrastructure for African AI. CRADLE's federated architecture provides the sovereign data infrastructure on which African AI models for education will be trained — continent-scale learning data under African governance, available for research and AI development without extraction (see Essay 12, AI in Africa's DPI-Ed).
Adoption network effects for V&P_Core. Every country that joins the federation adds to the continental intelligence, creating a network effect that strengthens V&P_Core's value proposition to prospective adopters. The data becomes more valuable — and the case for adoption more compelling — with each new participant.
Continental education intelligence for the EdTech Task Force. The AUDA-NEPAD EdTech Task Force requires evidence to coordinate continental education strategy. CRADLE provides the data infrastructure that replaces anecdote with evidence.
CRADLE follows the Breakthrough System's loosely-coupled governance model (see Essay 7, Section 5):
AUDA-NEPAD provides institutional coordination, continental legitimacy, and policy guidance on cross-border education data governance, exercising the implementation support mandate conferred by Decision Assembly/AU/Dec.973(XXXIX).
Ministries of Education retain sovereign authority over their national education data and participate in the governance framework through the AUDA-NEPAD EdTech Task Force.
Africa CDC is proposed to provide technical and institutional guidance based on its experience building the Central Data Repository and the Continental Health Data Governance Framework. The possibility of initially housing CRADLE's continental data unit within Africa CDC infrastructure (as proposed in Essay 15, Section 6) is a research question to be resolved during Phase 1.
The Spix Foundation provides engineering capacity and project management, integrating CRADLE's federation layer with the RESPECT Platform's data interfaces.
The AUDA-NEPAD EdTech Task Force provides the coordinating function for cross-country policy alignment, data governance, and trust preservation (see Essay 18).
Three options for CRADLE's permanent institutional home will be evaluated during Phase 2: a dedicated unit within AUDA-NEPAD's operations; initial housing within Africa CDC, leveraging its existing Central Data Repository infrastructure; or a new AU-aligned institution. The Phase 2 sustainability transition plan will recommend one option based on programme experience.
CRADLE's total programme cost is $10M ± $2M over 24 months: $6M ± $1M for Phase 1 (architecture and prototype, Months 1–12) and $4M ± $1M for Phase 2 (validation and scaling preparation, Months 13–24), with a go/no-go gate at Month 12. Budget detail, comparable initiatives, and post-research operating cost estimates are provided in the CRADLE Database Project Plan.
CRADLE depends on V&P_Core deploying RESPECT in the six pilot countries and on BEINGS developing the EMIS Interoperability specification. The primary risks are delayed V&P_Core deployment in pilot countries (which would reduce the data available for prototype validation) and the BEINGS specification timeline (which gates the EMIS integration work in Phase 2). Dependencies, risks, and mitigations are detailed in the CRADLE Database Project Plan.
Africa's health data infrastructure was built over two decades — from DHIS2's first deployments through Africa CDC's continental federation. Education can learn from that experience and build faster. The architectural patterns are proven. The governance template exists. The Malabo Convention is in force. And Africa's DPI-Ed is generating the first standardized, curriculum-aligned education data in the continent's history.
CRADLE provides the federation layer that connects it all. The 24-month programme will determine whether federated education data is technically achievable and politically viable at continental scale. If it is, the architecture, governance framework, and working prototype produced over those 24 months will enable continent-wide education intelligence that no single country's data could provide. Researchers will discover cross-jurisdictional patterns. Policymakers will benchmark with evidence. Results-Based Finance will operate on comparative data. And every additional country that joins the federation will make the whole more valuable.
Health proved it can be done. Education is next.