Earlier this month Fairfield Dolan presented the Business Analytics Showcase to give current students the opportunity to learn how graduates have used the skills they mastered in class to solve real-world problems.
Many of our MSBA capstone projects, which are a culmination of a year or two of hard work, have gone on to be presented at conferences, published in academic journals, and even win prestigious awards.
— MSBA Program Director Philip Maymin, PhD
Hosted by MS in Business Analytics (MSBA) Program Director Philip Maymin, PhD, and Business Analytics professors Chris Huntley, PhD, and Jie Tao, PhD, the Business Analytics Showcase featured capstone projects from two recent program graduates and one MSBA candidate.
“Many of our MSBA capstone projects, which are a culmination of a year or two of hard work, have gone on to be presented at conferences, published in academic journals, and even win prestigious awards,” said Dr. Maymin.
Breanna Cocuzzo '20, MS’21, now a technology consultant for Ernst and Young, presented “Classifying Vehicle Accident Severity Using Imagery Data,” a project to make auto claim handling by insurance companies more efficient with the use of deep learning models, an area of interest for her since she participated in the Dolan Co-Op Program at The Hartford. Walking the audience through her process, she described how the self-built model she initially used had struggled with limited training data, and she instead settled on a fine-tuned transfer model which was both generalizable and domain-specific.
“Learning programs like Python, R, and SQL is important, but what’s been a big differentiator for me has been the way the MSBA program teaches you to approach problems and [apply] logic,” she told the assembled group.
“Predicting Song Popularity with Advanced Learning Machine Models” is a project with great potential for those in the music industry, said presenter Bryan Einstoss MS’21, currently in charge of process automation for an investor services team. “Given a data set containing song features and metadata, is it possible to accurately predict the popularity of a song with a multiclass or multi-label approach?” he asked. “Secondly, how do song features evolve over time within different genres, and how do these changes correlate with the popularity of genres?”
Einstoss sourced data from Spotify, then created a model that broke songs down according to features such as loudness and instrumentals used. Analyzing trends in these features “can make it easier to spot new talent, and can help producers fine tune a song in production to see how they can increase the chance of making it popular for playlist curators.”
The final presenter, Mark Drummond, has had many years of work experience in finance and business analytics, and joined the MSBA program to take a deeper dive into analytics. He is now interning at Synchrony and will graduate from the program in December.
The goal of his capstone, “An Investigation of Wearables, Illness, and Recovery,” was to create a multi-class classification that could be used by clinicians to monitor a patient’s health status, then use that data to predict when a patient might be getting sick. Using Garmin Vigo and Oura tracking devices and monitoring heart variability, stress levels, and energy level, clinicians are, for example, able to identify some people who contracted Covid-19, before they showed symptoms. The problem is that the process is very manual, and not always accurate or repeatable.
“My subject matter expert wanted two things: the validation of what they were seeing and interpreting on the dashboard, and an automated method to predict where the patient was headed,” said Drummond. “If we can predict that, the healthcare provider can establish a better protocol for the patient,” he noted. In other words, providers can use the data in a way that’s proactive rather than reactive.
Fairfield Dolan reports a 100% employment rate for graduates of the MSBA Program. The program is now available as a fifth-year program option for qualified undergraduates from all Fairfield schools or majors. Students learn the skills most sought after by employees, including Python, R, Tableau, SQL, and more, and are eligible for exclusive Co-Op Program positions with top global firms.
