← All work
Unganisha
Blood donor matching for Kenya, privacy first
Unganisha (formerly DamuLink) connects a patient who needs blood with compatible donors nearby. It is my own product.
It is built for a domain where a wrong answer is fatal and a leaked detail is a legal problem, so safety and privacy are enforced at the database level.
- 1,620
- towns
- 47
- counties
- 1,051
- health facilities
- 163
- tests
Why
When a patient needs blood, the search runs on WhatsApp and phone calls. Donors are found late or not at all, and medical details spread with every forward.
What it does
- A 126-row derived antigen compatibility matrix that lives in a database migration. A wrong match is impossible by construction, not by convention.
- Spatial matching with PostGIS across 1,620 towns, 47 counties and 1,051 health facilities.
- Privacy by architecture: public request pages carry no names and are never indexed, identity data is purged on a fixed schedule, and nothing tracks HIV status or diagnoses.
- A separate n8n instance for the back office that cannot reach donor data at all.
- 163 automated tests.
What I led
- Everything, from the data model to the Kenya Data Protection Act reasoning behind it.