The Data Integrator is an open-source, self-hosted platform that ingests multi-source water and climate data (satellite and gridded products, in-situ sensors, national databases), harmonizes it into standard formats, and exposes it through a documented query API, a dashboard, and an optional AI assistant. An agentic AI capability designs, tests, and maintains the ingestion pipelines themselves.
It is developed by mWater for the UNICEF WASH Innovation Hub as the reference implementation of the Open Source Data Integrator Platform architecture, with initial deployments in Angola and Madagascar.
The architecture is specified independently of this implementation. It describes components by responsibility and contract so that any vendor can implement it with their own technology choices; the reference implementation is one set of choices that satisfies those contracts.
| Document | Contents |
|---|---|
| System Architecture | Design principles, system context, components, data flows, security and deployment architecture, extensibility |
| Interoperability and Standards Mapping | The standards each interface conforms to, integration paths for national systems, Digital Public Goods alignment, conformance criteria |
| Integration Library Specification | The machine-readable definition format for data sources, the job contract integration code runs against, and library conformance |
Diagrams are Mermaid and render directly on GitHub.
This repository currently holds the architecture documents. The source code of the reference implementation, the integration library, and deployment documentation are published here with the proof-of-concept deliverable.