Projects

Identification of Health Status via Gait Analysis

This project used ground reaction force data collected during walking for individuals with parkinsons disease and healthy controls.
I used R to load in data & isolate key points of the gait cycle to determine cadence for each foot through the entire trail.
I used feature engineering to measure & categorize all participants based on speed, variability, and asymmetry (difference between limbs). I visualized these differences, ran a principal component analysis and correlation analysis for feature reduction.
Finally, I created several logistic regression models to classify a test set as healthy or parkinsonian. I created several models beginning with speed as a single variable but settling on a stepwise logistic regression with moderate predictive power after doing diagnostics.


News Sentiment Analysis & Stock Prediction

This project used NewsAPI, several NLP models, yFinance, and logistic regression to predict stock prices.
This project used the free tier of NewsAPI, which allows for capture of the most recent 30 days of newspaper data for 128 sources.
In this project I selected all business sources and chose all articles that mentioned Apple.
I used TextBlob and Vader to perform sentiment analysis on both the titles and descriptions of each article.
Using the yFinance package I imported hourly stock data for Apple to match the time period recorded for the newspaper articles. I combined both data sources, split into training and testing and uses the Scikit-Learn package to preprocess my data and run an logistic regression & random forest classifier on the data with the outcome of whether the stock went up or down.
As an extra note I used Scikit-Learn to predict the stock price using a Long Short Term Memory (LSTM) model.


Spotify add on for BPM selection

This project uses python to access saved songs, filter based on a target bpm and tolerance, and create a playlist in the user account.
This came from an idea of how music of different genres impacts adherence & learning with gait retraining in runners. A modified version of this code was used to make the spotify web app above, that code is also in the respository linked below!


Online Computer Part Prices with Web Scraping

This project used the beautiful soup package in python to scrape price data from a list of predefined computer parts. This is the original version of the ebay app above, so try out the app, and more code is available on my github!