Projects

In collaboration with teammates Toni Akintola, Rachael Thumma, and Louis Cornett, our team overlaid the EJScreen dataset with the EJI index tool dataset from ATSDR, narrowing analysis to 16 key environmental and demographic features. We then:

• Applied a gradient boosted tree model with SHAP explanations to each census tract to identify the most prevalent factors contributing to cancer and respiratory risk.
• Built a pipeline that fed model results and demographic data into an LLM constrained to the NCSL Environment & Natural Resources Legislation Database to generate state-level policy recommendations.
• Created an interactive web application featuring a zoomable heatmap of the entire United States, allowing users to explore the prevalence of each risk factor at any geographic scale and generate custom policy proposals on demand.
• Won the Best Data Storytelling Award recieving $1000 in the competition for effectively communicating complex environmental health data and policy insights.

Abstract Overview Video Slideshow Team Picture GitHub Repo

This project focused on analyzing psychological techniques used in online scam messages. Allie Britton and I developed an autonomous machine learning pipeline in Python using Natural Language Processing (NLP) and a Multi-Layer Perceptron (MLP) model to detect these techniques.

The pipeline was integrated into a local full-stack web application using Flask, HTML, CSS, and JavaScript, allowing users to input scam messages and receive a detailed breakdown of the techniques used.

A training and validation dataset was manually built from 760 real-world scam messages to ensure accurate predictions.

Research Poster GitHub Repo

This project involved processing and analyzing large-scale retail data as part of the Data Club of Notre Dame. I worked with over 20 million lines of raw data using Pandas for cleaning and preparation.

In collaboration with 10 teammates, I developed a Tableau dashboard analyzing Kroger’s sales and customer segments. The dashboard included components to evaluate coupon campaigns, providing actionable insights into their influence on sales trends.

Slideshow Tableau Dashboard GitHub Repo

This project focused on improving the Amazon Prime Student membership experience as part of the Student International Business Council (SIBC) STEM program at Notre Dame.

As a team member, I conducted market analysis to identify weaknesses in the current implementation and researched the use of Generative AI lifestyle imaging to increase advertisement conversion rates.

Our team delivered a presentation at Amazon headquarters to over 100 employees, including the Head of Prime Marketing, providing actionable recommendations for future implementation.

Presentation

This project involved designing and building an interactive website as part of the Principles of Computing course at the University of Notre Dame.

I created the website using HTML and CSS to display various Python-based visualizations on global life expectancy trends from 2000 to 2015. Users can explore insights on health outcomes across continents through the interactive interface.

Visit Website