Feb 20 Post Update

Project Update: Switching from NBA to MLB Umpire Bias

After careful consideration and consultation Dr. Plancher, I have decided to switch my research proposal from NBA-rest-focused to MLB umpire bias. The new project aims to investigate, categorize, and utilize biases among MLB home plate umpires (pitch type, location, arm angle, release point, weather, umpire crouch height, etc.).

The decision to switch from NBA to MLB was driven by several factors. Firstly, the NBA rest proposal was a bit hard to categorize in a way that I found satisfactory (each model I though of contained some decently-large gap in coverage), and while there is still much to learn about the league, I felt that I would be more likely to make a unique contribution to the field of baseball by investigating umpire bias in MLB. Additionally, as a bigger fan of baseball, I am passionate about the sport and eager to learn more about its inner workings.

Over the past week, I have spent most of my time redefining the variables for my research and working on web-scraping the necessary data. The new variables are all derived from game context within the ESPN website. Web-scraping has been a significant challenge, as the data is spread out across various websites and formats (assuming I want the pitch location and movement data to be 3D), and must be cleaned and organized to be useful. I have used a combination of Python libraries such as Beautiful Soup and Pandas to automate the process and ensure that the data is accurate and reliable. This has required a significant amount of trial and error, but I am confident that the end result will be a high-quality dataset that can be used to answer my research questions.

While I have made progress on the technical aspects of the project, I have also spent some time thinking about the broader implications of my research. Umpire bias is a controversial and sensitive topic, and it is essential that my study is conducted in a fair and unbiased manner (although I much prefer human umpires to AI systems in the recent debate). To ensure that this is the case, I will be following a rigorous methodology and seeking feedback from other experts in the field (via reaching out to front offices and praying for responses).

Right now, I just have to finish up the data collection and then I can begin working on creating the ML model for the analysis.