Farm forecasting with real data and AI insights.

Foragecaster is 4-year research project with leading-edge industry and academia working on creating an accurate livestock supply forecaster using the best of breed domain models with machine learning. As a part of the project, I designed research-driven interfaces that translated complex data into practical, user-friendly tools for farmers. By conducting user interviews and collaborating with research teams, I ensured that advanced models were acessible and actionable for effective farm managament. ‍

In agriculture, accurately predicting pasture growth is essential for managing livestock and ensuring sustainable land use. Farmers need reliable insights into forage availability to make timely decisions on grazing, feed supplements, and herd management.

Without accurate predictions, they risk overgrazing, underfeeding, or having to deal with feed shortages, all of which impact farm productivity and profitability. The challenge lies in the complexity of modeling factors such as weather variability, soil conditions, and plant growth stages, which all influence pasture growth.

Foragecaster is a predictive tool designed to give farmers a clear, data-driven forecast of pasture growth rates over upcoming months. It uses a combination of weather data, soil conditions, and pasture types to provide predictions that reflect both expected growth and uncertainty.

By presenting information visually through interactive graphs, Foragecaster allows farmers to assess pasture availability, make informed decisions on livestock management, and better plan for seasonal changes. This tool provides a user-friendly interface that simplifies complex data, making it accessible for those without a technical background.

In this project, my role involved research and the design of the ForageCaster interface. I was responsible for creating a responsive, intuitive interface that would effectively communicate complex data in a clear and accessible way.

This involved designing interactive graphs to display predictive growth rates and uncertainty intervals, as well as integrating user interactions for filtering and customizing data views. Throughout the project, I collaborated closely with farmers over interviews to ensure the interface met user needs, iterated based on feedback, and tested the tool to ensure accuracy and usability.