This project utilizes red and white wine data gotten from UC Irvine's database, along with a rating done by judges on a 0-10 scale, 0 being the worst with 10 being the best. That data is then used to train a multinomial logistic regression model that will predict the quality of a wine data entry on that 0-10 scale. Once the model has done its predictions the accuracy is tested based on micro and macro ROC AUC scores that are then displayed.