dteuscher1 Goto Github PK

adjusted-plus-minus

Adjusted Plus Minus Models for WNBA players from the 2019 season. Adjusted Plus Minus (APM) and Regularized Adjusted Plus Minus (RAPM) models were fit providing an all in one player value metric for the WNBA.

catcher-framing

Project for generalized linear models class at BYU. Modeling the probability that a pitch is a strike using generalized additive models and determining catcher framing abilities and umpire influence on the strike zone using generalized linear mixed models.

forest-cover

EDA and model fitting for Forest Cover Type Prediction Kaggle Competition. Final results around the 69th percentile (top 31%).

irrigation_neuralnet

mlb-bayesian-ridge

Analysis of MLB pitchers and what factors influence a pitcher's FIP. Bayesian Ridge Regression is used to determine which covariates are most important. The model is fit using both hand written MCMC algorithms and Stan.

mmresults

movie-revenue

EDA and model fitting work for the TMDB Box Office Prediction Kaggle Competition. The goal of the competition is to predict the amount of box office revenue for a movie. Would be in the 58th percentile (top 42%) in the competition, if the competition was open still.

nba-basketball-visualization

Visualization and model fitting for Kobe Bryant shots over the course of his career. Data comes from the Kobe Bryant Shot Selection Kaggle Competition

nba-fantasy-basketball-app

ncaa-coaches

nlp-disaster-tweets

Natural Language Processing project for the Kaggle competition: "Real or Not? NLP with Disaster Tweets". Currently 51st percentile (top 49%) in the competition.

pitchclustering

portfolio

Source code for my website, which is a portfolio of most of my data science work

project426

A collaborative data science project done in Python analyzing the impact of hustle stats on the outcomes of NBA games. Overall, a logistic model proved best at determining game outcome. This project is for the Stat 426 class at Brigham Young University.

russiantopicmodeling

stat426-fall2021.github.io

Class repository for Stat 426 at BYU. Includes blog posts written by the students and the group projects completed at the end of the semester. The website can be viewed at: https://stat426-fall2021.github.io/.

store-demand-time-series

EDA, visualization and model fitting of time series data for the Store Item Demand Forecasting Kaggle competition using the Prophet package. The goal of the competition is to predict the sales of 50 different items at 10 different stores over 3 months. Results would be in the 57th percentile.