Projects


Pre-Snap Motion

Analysis of Pre-Snap Motion in the NFL

This project was completed for my Data Wrangling Final in the Fall of 2023.

A growing scheme that many NFL playcallers are utilizing is the concept of pre snap motion. Identifying coverage types, adding an extra run blocker, and momentarily freezing up a defense are all goals for coaches when using this strategy. The question is, how well does pre snap motion actually work in terms of efficiency and effectiveness? This past semester I have been conducting research to find the change in yards per play, redzone success, the difference in separation, and others trends when comparing motion plays to non motion plays in the NFL. The data and trends found will be used to answer the following three questions:
1. Does motion impact the effectiveness of yards gained each play? What teams utilize the most motion in the NFL?
2. Does pre-snap motion have more success in situations like 3rd down conversions and red zone opportunities? Does motion usage rate change depending on the situation?
3. Is there a correlation between teams that utilize motion the most and the most successful offenses in the NFL?
The report includes the data used, the procedure of my analysis, and the conclusions drawn from the research questions.

View Project Report

Pre-Snap Motion

NCAA March Madness Seeding Analysis Model

This project was a group effort completed for my Data Mining Final in the Fall of 2023.

Every March, the NCAA Tournament Selection committee faces a dilemma of who to put into the field, as well as what the seeding of the field should be. I helped create a model that can assist the selection committee to make seeding decisions and which bubble teams are in and out for future NCAA tournaments. A dataset sourced from Kaggle.com provides many useful performance metrics for every single team starting in 2014, excluding the 2020 non-tournament season. Notable features from this dataset include wins, strength of record, shooting percentage, rebound percentage, and turnover rate. The response feature for this model was labeled “MakeTourney”, a binary variable used to determine if a team qualified for the tournament that season. Logistic Regression became the clear model of choice for this project. The chosen model had the strongest AUC and the best confusion matrix distribution when compared to other model types.

View Project Report

Capstone

Iowa City SSMID Capstone Project

This semester our capstone group was responsible for updating dashboards and providing analytical insight to the South of 6 Self Supporting Municipal Improvement District. This process was started last summer, and a large portion of the project was spent updating the data they were using. In addition, our group implemented some new data sources to showcase including a community impact survey and location traffic data. The final deliverable was a polished final dashboard that the South of 6 community can use in order to gain insights into traffic, engagement, and feedback to help improve the district.

View Project