Health information systems · DHIS2 · interoperability
Seven real projects across native DHIS2 apps, DHIS2–FHIR interoperability, and applied statistical outbreak detection — each built from a real gap, a real published method, or a real reported bug, and live-verified against real instances and real data, not demos.
Real projects
7
Technical write-ups
3
Synthetic demos
0
Live-verified
All
Projects
Every one built end to end: real API mechanics confirmed live before shipping, real test suites, real bugs found and fixed before release.
DHIS2 native app
Evidence, coverage, freshness, and RDQA-aligned quality reporting for eight disease surveillance programmes, computed live from DHIS2's own API.
DHIS2 native app
Admin-configurable, RDQA-aligned quality checks for any dataset on any DHIS2 instance — no bundled programme list, including AMR-domain presets grounded in WHO GLASS.
DHIS2 native app
Guided, secure external data sharing: scoped CSV export or a revocable API account, with the DHIS2 personal-access-token boundary handled honestly rather than papered over.
DHIS2 native app · Tracker
A WHO AWaRe-aligned antibiotic prescribing checklist. Watch/Reserve prescriptions require a justification note before submission, with a compliance summary flagging entries that bypass it.
DHIS2 native app · Tracker
A supportive-supervision checklist built directly from Burnett et al. 2019's published evaluation of PATH's MalariaCare Electronic Data System — tracking completeness and competency as two separate, honest scores.
Interoperability · Node CLI
A real DHIS2–FHIR interoperability tool, not another app-shell demo: syncs Immunization resources from a live FHIR R4 server into DHIS2 Tracker, with verified idempotent re-sync.
Applied statistics · Node CLI
An implementation of the Farrington outbreak-detection algorithm (1996), validated against the UK's full real COVID-19 case history — correctly flagging every documented wave, including Omicron.
Writing
Longer-form pieces — what building this work actually taught me, and two real methods checked against real data.
Retrospective
Seven platform behaviors the docs don't mention, each found by an actual account clicking the actual button.
Read →
Visual analysis
The full real numbers from Burnett et al. 2019's evaluation of PATH's MalariaCare Electronic Data System, visualized.
Read →
Visual analysis
The Farrington algorithm's alarms, laid directly over the UK's real COVID case curve.
Read →