Project Spotlight: Andrew Warburton
Last semester, Forge student Andrew Warburton took Node, our data science Skills Course, and his final project was voted best Node project in our 2021 Fall Showcase! We asked him a few questions about his project and his experience with Node. To see all the Node projects from last semester (including Andrew’s!), head to joinforge.co/fall21showcase-node.
#1 Tell us a little bit about yourself and your history with Forge.
“I graduated from York College of Pennsylvania with a bachelor’s degree in Engineering Management. Following graduation, I worked for Worthington Industries (NYSE: WOR) as a Transformation Manager before joining PwC (now Guidehouse) as a Strategy and Operations management consultant. I chose to come to UVA to earn my master’s in Systems Engineering and gain an advanced level of understanding of the field of management science.
While attending classes at UVA, one of my professors mentioned Forge, which led to my discovery of Node. Because data science is a fundamental component of management science, I felt that participating in Node would be a great opportunity to supplement my coursework with practice in an area relevant to my interests and goals.”
#2 What was your favorite part about Node?
“I enjoyed the flexibility that Node offered the most. It let me apply data science techniques to a variety of subjects that I found interesting. Additionally, with its casual and laid-back atmosphere, Node provided the opportunity to explore interesting subjects without the rigorous pace and associated pressures that often comes with delivering work in professional or academic settings.”
#3 Tell us about your project, what inspired you to make it, and how you created it.
“I built a machine learning model to predict the outcomes of the current season of the English Premier League (EPL) using football/soccer team statistics. Every season, the top four teams in the EPL qualify for the Champions League, which is essentially the EPL’s playoffs that determines a league champion. My model was designed to predict which four teams would make it into the Champions League and separately predict which team would become champion.
The model that I created would make these predictions by reading various team statistics, such as the percentage of shots a team took from a given area on the field or a team’s through-ball passing accuracy, to identify patterns associated with previous teams that qualified for the Champions League and/or became champions. Next, I applied this model to look for similar patterns in current-season statistics, using the information to predict which teams were most likely to qualify for and win the Champions League.
The most challenging (or tedious) part of the project was synthesizing all the data for the model to analyze. I collected EPL statistics from numerous sources, ultimately using 27 separate datasets for the project. All this information had to be transformed into something cohesive for the model’s predictions to be useful and as accurate as possible, which took a decent amount of time. Ultimately, this effort proved worthwhile because the predictions made by my machine learning model mirrored the predicted outcomes for the EPL’s season in sports betting apps, such as DraftKings. These organizations are known data science powerhouses, and it was rewarding to see my results reflected in theirs.
I enjoyed working on this project because I had the opportunity to apply statistical/data science techniques to a topic that I’m interested in. For some extra fun, I placed some small bets on the outcomes of the current EPL season that were predicted by my model. If they work out, it will be nice to say that my machine learning model made me some money!”
