The overall project goal is to train a machine to predict a musical genre given any album cover. This will be implemented using several ML techniques covered in Stanford's CS231n such as using a knn classifier, linear classifier, and cnn's. The first goal is to acquire the dataset needed to train such a machine. After doing some exploring on Kaggle, it seems that there doesn't exist a dataset consisting of album covers along with their associated genres. Thus, I've decided to create my own by scraping data off of AllMusic's advanced search page.