This is a collection of personal projects I have done related to data science and machine learning. You can find all of these projects along with the relevant code on my GitHub.
Predicting NFL Play Calling Success
I performed a classification analysis on a data set of NFL play-by-play data to predict the success of a rushing or passing play in a given game scenario. I used three total algorithms to model the data: random forest, boosted random forest (XG Boost), and a neural network. My algorithm exhibited a maximum of 57% precision as compared to 49% inferred precision by NFL head coaches.
Used Car Price Prediction
I performed a regression analysis on a data set of used car prices to search for future bargains. I used four total algorithms to model the data: decision trees, random forest, k-Nearest neighbors, and a neural network. I achieved consistent accuracy of my price predictions within 25% of the true listing price.
Predicting Stock Market Moves Using a Neural Net
I built a package that implements a neural network to predict increases/decreses in stock prices. The network implements a strategy that consists of making lots of small trades in the hope that lots of small wins leads to large total gains. I achieved a 92% accuracy on S&P 500 stock predictions.