Portfolio

Data Science Projects

NFL Expected Points Analysis

Expected Points is an equivalence metric for American Football that has been around since the 1960’s and was reborn in the 2000’s. In this project, I utilize the nflfastR data set, along with the ggplot2 and tidyverse packages to perform a regression analysis on the the data to determine what statistics most highly correlate to Expected Points Added and look at any predictive value of the models.

View Code on GitHub

EPA as a Predictor

In the National Football League, the Quarterback is the single most important position. Not just in its own sport, but no other modern team sport is as heavily dependent on a single position like American Football is on the Quarterback. In this project, I continue my analysis of Expected Points by comparing it to other modern metrics and utilize multivariate regression to determine their predictive power. As well, I created an RShiny app to show the correlations between the metrics.

View Code on GitHub

Exercise Prediction

Using devices such as Jawbone Up, Nike FuelBand, and Fitbit it is now possible to collect a large amount of data about personal activity relatively inexpensively. In this project, I will use data from accelerometers on the belt, forearm, arm, and dumbbell of six participants that were asked to perform barbell lifts correctly and incorrectly in 5 different ways. Using this data, I attempt to correctly predict which exercise the participants were doing and if they did them correctly or not.

View Code on GitHub

Data Visualization Projects

PFF 2021 College Coverage Statistics

This project was done using a combination of RStudio for the exploratory data analysis and Tableau to visualize the data. I chose to look at 2021 college football statistics for coverage players provided by Pro Football Focus.

NFL Exploratory Visualizations

This project was done to show how to explore and analyze data through visual means without using a dedicated visualization tool. The data is from the nflfastR package and uses visualizations to answer multiple questions.

View Code on GitHub

Algorithms

The following are Python projects of Machine Learning algorithms written from scratch (no ML packages used).

Skilled-based Projects

A selection of smaller projects utilizing various DS and ML skills and packages.

Page template forked from evanca