FoodMuse: Health-Aware Food Recommendation System Using Machine Learning and API-Based Food Data

FoodMuse is a meal recommendation system created to assist users in eating healthier and in-turn, reducing food waste. The system will provide recommended meals based on the user’s health condition and sustainability goals by taking into consideration the appropriate health restrictions (e.g., diabetes) and ingredients that are available to the user. Built with a React.js for front-end and a Flask for back-end, FoodMuse make use of the Spoonacular API for recipe retrieval and uses Random Forest (RF) to evaluate nutritional suitability, achieving F1-score of 0.8292 and accuracy of 0.881 -outperforming Decision Tree (DT) and Logistic Regression (LR) models. In order to further improve functionality, IndexedDB provides offline access to saved meals, while Gemini AI creates new recipes from leftover dishes. FoodMuse makes meal planning simpler, healthier, and more sustainable by combining food data, machine learning, and a user-friendly interface.