Food Data Nutrition Analysis

Collaborators: Daniel Duan, Yuxuan Gao

In this project, we conduct an analysis of food nutrition using the Open Food Facts dataset provided by Kaggle! This dataset contains nutrition facts about food products from all over the world. First, we spend some time cleaning up and reformatting the raw data while analyzing it for any interesting observations or correlations. Next, we focus on using and testing multiple models (Decision Tree, Logistic Regression, Linear Regression, Random Forest Regression) to see if we can accurately predict a product's nutrition score based on major ingredients in the product. We also used computer vision to attempt to train a model to classify foods as "Very Healthy," "Moderately Healthy," "Moderately Unhealthy," or "Very Unhealthy" based off of their images. Lastly, we conduct a brief regression analysis on how much energy food items provide.