DHIS2 Implementation Adviser & Data Scientist

Brian
O'Donnell

Hi! I'm a public health data specialist with 10+ years of experience building durable national health information systems. I also write code to help unlock these data for deeper exploration. This is a portfolio of all my projects in health data analysis and software development.

Brian O'Donnell

00 — Portfolio Timeline

Eras of Work

01 — DHIS2 Apps

Custom Applications

React applications built on the DHIS2 App Platform for visualizing health data, exploring tracker events, and surfacing program insights.

DHIS2 App

Trend Sketch

User-generated predictions of trends in DHIS2 indicators to enable crowd-sourced forecasts.

DHIS2 App

Event Relationships

Capture app plugin to store event relationships within a patient's DHIS2 record.

DHIS2 App

Program Indicator Boundaries Explainer

Interactive visualization of time dimensionality in indicators to support configuration.

DHIS2 Metadata

Health Data Toolkit

Collab wwith WHO on model DHIS2 systems for immunization, HIV, TB, maternal health, etc

02 — R with the DHIS2 API

Analysis & Visualization in R

Using R to connect, extract, and analyse data from DHIS2 — from network graphs of COVID surveillance to tracker data workflows.

R · igraph · Shiny

Covid Networks

Mapping COVID-19 transmission networks via DHIS2 case surveillance data.

R · Shiny

Program Indicator Generator

Automate DHIS2 program indicator creation in bulk with the metadata API.

R · RMarkdown

User Activity in Tracker

Analysing DHIS2 user login patterns and dashboard engagement via the API.

R · ggplot2

Visualizing Patient Movement

Experimental diagrams of patient paths in MCH programs.

R · ggplot2 · Tidytext

Statistics in West Wing Dialogue, 2020

Early data mining project on the language of statistics in The West Wing.

03 — Epidemiology

Research & Analysis

Academic and applied epidemiology: predictive patient attrition models in Nigeria and reproductive health monitoring in Palestine.

Prediction · Hypertension

Masters Thesis: Hypertension Dropouts

Predicting dropout from care among 40,000 hypertensive patients tracked by DHIS2 in Nigeria.

Epidemiology · DHIS2

Maternal Healthcare in Palestine

Built SMS&dashboard pipeline for eRegistry; analyzed data for publication.

R · Survival Analysis

Survival Curves

Kaplan-Meier and Cox regression models for time-to-event outcomes in health data.

04 — TidyTuesday

Data Visualization Portfolio

Weekly data visualization challenges in R. A growing collection of #TidyTuesday submissions exploring ggplot2, data storytelling, and visual design.

2020-05

Cocktails Prices in Norway

2020-06

Slaves Freed During Transatlantic Transport, 1860-1870

2020-06

Avengers and the Bechdel Test

2020-08

Spacewalks Networks

2020-10

US Beer Awards

2021-05

Broadband Access in Austin, TX

2021-08

Afghan Refugee Resettlement in the US

05 — Blog

Writing & Research

Long-form pieces on public health data, heat-health early warning systems, and cancer registry infrastructure.

Mar 2026 Heat Health Early Warning Systems — country web search Feb 2026 DHIS2 use in cancer surveillance — a global landscape