Fertilizer Recommendation ML Tool.
Through a wide variety of mobile applications, we’ve developed a unique visual system and strategy that can be applied across the spectrum of available applications.
Meta:
- Organization: Alliance of Bioversity International and CIAT
- Role: Data Dashboard Engineer (Consultant)
- Duration: March 2023 – May 2023
- Tools: Python, Django, ReactJS, PostgreSQL, Machine Learning
Short description: A machine learning tool that tells farmers in Tanzania whether fertilizer or alternative agronomic practices will actually improve their crop yield.
Full body: Fertilizer isn’t always the right answer — sometimes better agronomic practice does more for yield, and getting that recommendation wrong costs farmers money and effort. Using agronomic survey data from Babati District, Tanzania, I built an interactive machine learning tool that predicts whether a farmer should apply fertilizer or adjust practices instead.
The build was full-stack: a Django backend serving the ML model’s predictions, a React frontend farmers and field staff could actually use, and a PostgreSQL database I designed to support it, complete with API endpoints connecting the two layers. It’s a small-scope project by budget, but a clean example of taking a model from research question to something a non-technical end user can open and act on.
