Code
I have a variety of projects on GitHub. You can view them right here. Some repositories contain more recent personal projects and some contain assignments from my undergrad and graduate classes.
My personal projects include this website! And also:
- Goats Stats: I used sentiment polarity scores for song lyrics by the Mountain Goats to classify albums as belonging to their lo-fi or hi-fi period. I did this for both word-tokenized and sentence-tokenized lyrics from 16 different albums. I tested four different machine learning algorithms: Naive Bayes, Decision Tree, K-Nearest Neighbors, and Random Forest.
- Creators Thrift website, which can be found live right here: https://www.creatorsthrift.com/. I donated my time to work on this website, which seeks to unite artists from various disciplines who might collaborate or exchange skills.
- Actor or Movie?: Classifying Wikipedia Articles: I used part-of-speech and named entity recognition tagging on word-tokenized Wikipedia pages for 10 actors and for 10 movies. Using counts of the most common POS and NER tags per actor and per movie, I was able to classify the articles as belonging to the Actors category or the Movies category. I tested three different maching learning algorithms: Naive Bayes, Decision Tree, and Gradient Boosting.
- Cage Data: I curated a Nic Cage data set using information from IMDB and Rotten Tomatoes. The initial dataset was updated in January 2022. Both the initial and updates .csv files can be downloaded from Kaggle: Nic Cage Movies.
My school projects that I am most proud of are:
- In grad school, I worked on a project I named Polls and Polarity that compared the political opinions of those who take opinion polls and those who tweet about politicians. The opinions of twitter users were determined using sentiment polarity analysis on tweets mentioning certain politicians. The specific politicians studied for this project were the three front-running Democratic candidates in 2019: Joe Biden, Elizabeth Warren, and Bernie Sanders. Both the opinion polls and tweets formed a time series, with tweets gathered from the same dates as the opinion polls. This was done in an effort to form a contemporaneous view of how the two sources of opinions might change over time. Ultimately, I found that the opinions were completely distinct. A visualization of the project can be found at this link. This project was completed as the first practicum project for my MS in Data Science from Regis University.
- In undergrad, I worked on a hi-fi mockup of my idea for a redesign for Mendley desktop. I had worked on this project with another student for the Human-Computer Interface Design class (COMP 441) at Loyola University Chicago. I really enjoyed learning about making an interface intuitive for the user.