Detecting Bias in University Admissions

Technologies: Figma, TypeScript, Python, Tailwind CSS, Next.js

Our senior design project focused on detecting bias in university admissions. The project had two main aspects: building a web tool to flag biases and a human study portion to offer insight into trends we see in the data.

We developed a web app that takes in a student dataset and runs a suite of metrics to determine statistical bias. We use the university's region as a baseline that we use to determine if a discrepancy is significant enough to constitute possible bias. We then display the results on a dashboard for the user to view.

After we ran UT's data through our tool, we moved on to the human study portion of our project. The goal of the human study was to present our findings and get our interviewee's thoughts. We also wanted to know about actions student organizations or the university are taking to address some of the discrepancies we observed in the data.

I was the project's team leader, and I worked primarily on the website's frontend and on conducting the interviews for our human study.