Developing a Shopify web application that provides customized administrative functionalities (e.g., complex queries, advanced scheduling features, custom data fields) for a startup flower delivery company, with a focus on supply chain optimization and an aesthetically accessible and multilingual UX/UI design.

Tech specs: Shopify API, GraphQL, JavaScript, Liquid, Prisma, SQLite

Completed under the guidance of Dr. Richard Zemel.

Implemented several recommender system (RS) models via different recommendation algorithms/strategies (e.g., myopic, exploration-promoting) to simulate long-term user behavior and RS ecosystem dynamics.

Conducted comparative analyses to investigate the interplay between RS metrics/dynamics, including accuracy, recommendation quality, user behavior homogeneity, and item fairness.

Developed a website using Jekyll.

Final research report: Bubble Bursting Implementations and Dynamics in Recommender Systems

• Developed a machine learning model to select representative high-frequency data cycles based on Connected Vehicle metadata to predict customers’ vehicle charging behaviors and patterns, written in Python and HiveSQL.

• Performed data wrangling on over 50,000 unique vehicle identification numbers.

• Evaluated the UX/UI of several internal web applications to improve the efficiency and user experience of employee queries for Connected Vehicle data.

• Managed and trained a team of four undergraduate research assistants.

• Conducted and supervised an audit of how popular news outlets use visual representations to display wealth distribution and inequality data, and how these representations perpetuate systematic bias in viewers’ perceptions and assessments of wealth inequality.