Projects

Restaurant Review Dashboard

Because restaurants are a customer service business, insights in how a restaurant is performing can be gained by observing customer sentiments in online reviews. Reviews for a restaurant were scraped off the web using Python’s scrapy, processed to show meaningful information, and displayed in a Tableau dashboard. The dashboard is able to clearly show recent monthly performance in terms of customer ratings as well as common keywords in their reviews all in one place.

Anime Recommender

Item-based collaborative filtering was done on user rating data that I scraped off of myanimelist.net. Big data issues due to the sparse nature of user rating data was dealt with by making use of compressed sparse matrix formats to perform calculations. Recommendations for a user were made based off of a user’s watch and rating history combined with the similarity score values found from all of the user rating data. A demo of a web app to recommend animes can be seen at here or by clicking on the project title for an embedded view. To see recommended shows for the user dektuh, type “dektuh” in the username username section.

Features that Affect Anime Rating and Ranking

This project focuses on exploring what features about an anime tended to perform better. The analysis looks into genres, voice actors, production studios, and media type. The goal was to gain insight into what the public was interested in and what they ultimately enjoyed.

Mammogram Mass Classification

Mass data was used to classify the severity of the mass through different machine learning models. The models accuracy, precision, and recall score were compared to make a conclusion about the best performing model.

Contact Info

Resume
Email: dexkluu@gmail.com
Phone: 510-910-0935