About Participate Hackathons Main Stage 2025

DigiEduHack 2025

Rethinking education in the age of digital skills.

NOVA prediction

A solution proposed for the challenge The Food Challenge

Solution details

We propose a web platform that explains the NOVA food-processing classification system and provides an AI-powered tool that predicts the degree of processing for food products using Open Food Facts Canada data. Its main elements include clear educational content adapted from the "Guia Alimentar para a População Brasileira" and foundational NOVA publications on ultra-processed foods, an AI-generated FAQ, and a Python-based ML model that estimates processing levels for products lacking this information. The first milestone is to build the basic website in one day, followed in the subsequent weeks by data cleaning, feature engineering, model training, and integration of predictions and explanations. Resources required include a team of 4 students, a supervisor, computing resources (CPU/GPU optional), and the Open Food Facts dataset. Barriers may include incomplete product data, ambiguity in ingredient lists, and model performance constraints. The solution enhances digital education by turning a complex classification framework into an accessible AI-supported learning tool, and its success can be measured through user engagement, prediction accuracy, platform usage, and feedback on clarity and usefulness.

Tweet / Slogan

From data to insight: our AI platform explains NOVA and predicts food processing levels, supporting students, educators, researchers, and consumers in understanding how processed their food really is.

Back to The Food Challenge