05
Web
2024

RwandaHealthObservatory

A national health data visualization system tracking WHO-categorized disease indicators from province to cell level across Rwanda.

Role

Frontend Engineer & UI Designer

Year

2024

Type

Web Platform

React
TypeScript
D3.js
Figma
Custom Data Viz
The Problem

The Rwanda Health Observatory system was built to track and visualize the indicators of diseases categorized by WHO. There were different diseases to track — but the hardest part was maintaining the sub-causes. Each disease had to be categorized within its sub-causes, and this had to be visualized from province level down to cell level.

The data also had to change dynamically as the visualizations changed selecting a province had to update everything, drilling into a district had to update everything else.

The Solution

A unified health dashboard where officials can navigate the full indicator hierarchy, apply geographic filters from province to cell level, and see how sub-causes relate to their primary disease categories, all visualizations updating simultaneously.

We also needed to visualize for all parts of the country and change the dynamics of the data as the visualizations change.

The Approach

I worked on both the frontend implementation and the design. The hardest part of the UI integration was getting the logical placement right between the disease indicators and the sub-cause indicators making that relationship self-explanatory without documentation.

The placement of related indicators had to form categories based on how common data or subjective causes relate to their major disease name. I was focused on making it integrate into a logical base and a flawless understanding of both the metadata of the indicators and how they can be related together.

The Outcome

Health officials can now track indicator relationships and geographic trends that previously required hours of manual work. The interface was built to be usable on day one without training documentation which for a government health system serving non-technical officials is the real measure of success.

Iterations
01

Redesigned the sub-cause connector after users could not distinguish between parent and child indicators in the initial tree view

02

Added a freeze province toggle after officials kept accidentally deselecting their geographic filter while drilling into sub-causes

03

Rebuilt the choropleth color scale to use relative rather than absolute ranges, outlier districts were making the rest of the map unreadably uniform