To explore unsupervised learning and NLTK, I wanted to base my project off of one of my favorite things in the world... FOOD! As the former editor of my college newspaper's food section, I'm fascinated by how people talk about food and how people describe good food versus bad food. This project used natural language processing on Amazon Fine Food reviews to reveal which types of food and which qualities of food score the highest reviews to give insight on which products Amazon should buy more.
I used a K-Means classfier to determine the groups that five-star Amazon Fine Food reviews fall under. After trying many cluster numbers, I determined that five clusters provided the most coherent groups.
This folder contains the following files:
- Presentation slides (five_star_slides.pdf)
- Code (project_five_star.ipynb)
Click here for the accompanying D3 visualization!