Master’s in Applied Data Science Summer 2024 Capstone Winners
The Capstone Project is a cornerstone of the MS in Applied Data Science program, pairing students with industry partners to tackle pressing data science challenges. At the recent Summer Showcase, a diverse range of projects showcased the innovative solutions developed by these talented individuals. Six teams were particularly recognized for their exceptional work.
Greg Green, Senior Instructional Professor and Director of the MS in Applied Data Science program looked back on the Capstone Showcase, stating, “The advanced AI and Data Science Capstone projects completed this summer spanned many technical areas including NLP, Computer Vision/Deep Learning, Advanced Machine Learning and Large Language Models just recently made available. We continue to be impressed not only with the advanced thinking within the technical solutions but the clarity of communication and presentation styles that make these solutions understandable to senior business leaders from a variety of domains of expertise.”
One of the Summer Capstone winners, Peerapak Adsavakulchai, observed that one key strength of his team was their diverse backgrounds. His teammates came from various fields, including trading, consulting, and data science-driven research. This diversity contributed to each team member bringing a unique perspective and skill set to the project.
Learn more about the projects that won Best in Show below.
FTI Consulting | FTI Chat
Presenters: Peerapak Adsavakulchai, Jason Rajan, Kanu Madhok, James McKenzie
Faculty Advisor: Abid Ali
Project Topic: This capstone group developed an AI-powered business development tool for FTI Consulting. FTI and other consulting companies, operate based on project-to-project basis and it is hard to maintain a traditional, and more efficient databases like other companies that conduct transactional businesses. A lot of the data resides in “inefficient” file types. FTI Chat, a chatbot leveraging LLMs, automates project approach creation based on client needs. FTI Chat retrieves past project data using RAG and generates responses via an open-source LLM. Due to its ease of usage, FTI Chat has the potential to become a staple product that is deployed independently where different teams can use to streamline their business development process.
Independent Student Project | Federal Government Contract Dollar Awards to Small Businesses for Fiscal Year 2023
Presenters: Zhengliang Jiang, Alexander Saucedo, Yansong Guo
Faculty Advisor: Anil Chaturvedi
Project Topic: For the last three fiscal years, the United States government has failed to meet
prime contract award goals for Women-Owned Small Businesses (WOSB) and Historically Underutilized Business Zone (HUBZone) businesses as established in the Small Business Act 15 U.S.C. 644(g). To understand the shared attributes of these two business categories and the business categories that met the award goals, this study analyzes 2023 federal contract awards using autoencoding models for dimensionality reduction and t-SNE for visualization to identify clusters. This team’s findings provide insights into which agencies and types of contracts were awarded to various small
business categories.
Tilia Holdings | Strategic Growth in Food Manufacturing and Distribution: Natural Language Processing Approach
Presenters: Kinjal Joshi, Jaelynn Kim, Benjamin Li, Sophia Fumagalli
Faculty Advisor: Gizem Aydin
Project Topic: Tilia Holdings, a private equity firm specializing in food manufacturing and distribution investments, lacks an in-house approach for market research. This team used natural language processing (NLP) to create a tool focused on analyzing earnings call transcripts of major food companies. This tool provides current and historical strategic focus, market outlook, and maturity and sentiment analysis. It aims to reduce Tilia’s market research costs, provide proactive insights, and enhance investment success.
Honorable Mentions
Ateema | AI-Driven Chicago Show
Presenters: Hyojun Kim, Marian Xu, Chui Ning Kung, Aiman Al-Taubi
Faculty Advisor: Sanjay Boddhu
CogniEdu | AI-Powered Academic Planning for Student Success
Presenters: Mike Meissner, Joe Strickland, Adela Cho, Roselyn Rozario, Forough Mofidi
Faculty Advisor: Nick Kadochnikov
UChicago Medicine Department of Radiology | Impact of Heterogeneous Model Ensembling on Detection of Prostate Cancer from MRI
Presenters: Devan Richter, Siobhan McDermott, Connor Mulshine
Faculty Advisor: Batu Gundogdu