This is an app I created to help with data collection for my Cadence study. It has a semi customizable timer used to altert the team when different buttons needs to be pressed as well as a counter to calculate walking step rate based on different time intervals.
This is an app I created to complete the data processing necessary for the biomechanical outputs I receive from the Cadence study. You can upload any number of text output files from our MotionMonitor system and this script will segment the walking data into steps, filter said steps based on a few criteria (ie. not crossover stepping on the instrumented treadmill), and create an output sheet with peak and impulse values for each limb during each trial. If you're interested in testing out this script contact me and I can provide you with a few practice files that weren't involved in the study.
This is likely the most useful application for the visitors to this website though is becoming outdated. This app allows you to explore the Fair Market Housing data for the 2024 fiscal year. Fair Market Rents are 40th percentile rents aggregated by the US Department of Housing and Urban Development. The primary functionality allows you to visualize these rent values for studio apartments up to 4 bedroom units within a set mile range from a zipcode. Additionally, you can complete a head to head comparison between two cities.
The projects in the below section are a mix of personal and academic projects that I have worked on over the years and have been of great use to me ranging from initial exploration of gait retraining, applications to find the best priced computer parts, and an analysis of the best submissions to learn in brazilian jiu jitsu.
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.
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.
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!
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!