Flagship work

A multi-agency early-warning bulletin system for Tanzania

A platform that lets several national agencies contribute their inputs and produces impact-based forecast bulletins — with the maps drawn, the documents formatted, and every change recorded.

TMAPMO-DMDMoWGSTMoHMoANEMC
Mount Kilimanjaro and surrounding cloud, photographed from the International Space Station
Northern Tanzania under cloud, from the ISS, July 2024 — the weather the bulletins are about · NASA

The problem

Impact-based early warning in Tanzania involves more than one agency: meteorology for the weather, water for floods, geology for ground hazards, health, agriculture and environment for what the impacts mean. Their inputs used to be assembled by hand into bulletins — slow, inconsistent, and hard to audit.

The outcome

What took a working group hours of manual assembly becomes a scheduled, reproducible output — and the same design generalises to any multi-agency reporting problem, from seasonal agricultural outlooks to flood situation reports.

PythonStreamlitFolium / Plotlypython-docxPostgreSQL

What was built

  1. 01

    A role-based web dashboard where each agency enters its own inputs on forms and maps

  2. 02

    Two bulletin products: a five-day severe-weather impact-based forecast, and a three-day multi-hazard bulletin combining all agencies

  3. 03

    Automatic map generation of Tanzania at region or district level, with alert colours, labels, water bodies and hazard icons

  4. 04

    Bilingual English/Kiswahili templates, generated as Word documents and converted to PDF

  5. 05

    Data bridges so one agency's inputs pre-fill the next stage, plus a browsable history and admin audit log

The same design fits any multi-agency report

Seasonal agricultural outlooks, flood situation reports, environmental monitoring bulletins. If several institutions feed one document, this is the shape.

Talk to us about yours