Last month, Fairfield Dolan’s Business Analytics Showcase highlighted three impressive projects that proved the breadth of use for analytics in various fields. The event was hosted by MSBA Program Director Philip Maymin, PhD, and Associate Professor of Information Systems and Operations Management Christopher Huntley, PhD.

MS in Business Analytics students Breanna Cocuzzo, Bianca Gadze, and Mark Drummond presented the three projects they collaborated on, bringing their different backgrounds together: Drummond has years of work experience, Gadze majored in math and economics as an undergraduate, and Cocuzzo was in marketing. All three will graduate from the master's program this spring.

The group’s first project, “Inmate Deaths,” examined whether many of the deaths in prison should be considered suspicious. Data collected on 7,570 deaths from 489 prisons over 11 years showed that 35 percent of those deaths were attributed to suicide and drug use, and half were attributed to natural causes. The remaining causes were either not reported or considered “other.”

"Compared to the general population, we found that inmates are five times more likely to have serious mental illness and two-thirds are more likely to have a substance abuse disorder,” noted Gadze. The group also found a concerning trend upwards in deaths attributed to mental illness, especially in the late summer and early autumn, which led them to ask whether guidelines for counseling were being followed.

The team also noted a disturbing amount of missing data, particularly among those classified as having a “natural death.” For example, 54.5 percent of these deaths did not record age. “We found we could organize our findings into two areas: suspicious deaths and suspicious data,” said Cocuzzo. “The term natural death needs to be accompanied by more information…it’s just used too often without a whole ton of context.”

“Anytime you have those kind of data quality issues, especially around something that is literally life and death, you have to wonder what else they're not measuring,” warned Dr. Huntley. When it comes to any business, the absence of reliable metrics will seriously impact the ability to produce accurate models that positively affect business decisions.   

The trio used Excel, R and advanced stats to gather, cleanup, analyze, and forecast their findings.

Next up: “Basketball in a Bubble: COVID-19 and NBA-20,” a deep dive into the NBA-mandated isolation bubble surrounding players, and the effect it had on performances over the season. In this perfectly controlled environment, the students asked, did players perform better?

The group used Pandas and Numpy for clean-up and preprocessing data, imputation and analysis, and Matplotlib and Seaborn for exploratory data analysis and visualization. DeepNote allowed them to work on code simultaneously. In order to create a consistent player set, the students controlled for athletes who played in three seasons (2018-2020), who played at least 50 percent of games, and who appeared in the playoffs. 

This narrowed the sample set to 11 players, who were then ranked according to their performances in both offense and defense. Data revealed that six of the players improved overall, while five performed worse within the bubble. And while the main focus was on analyzing performance, “we found that [most of] the top five performers did not have children or a spouse, and their average age was 27,” said Cocuzzo. “If you look at the bottom six, every one of them has a spouse and at least one child. Their average age was 32.” Further analysis into the social and emotional aspects of isolation would be interesting, Gadze noted.

The final project, “Investigation into Wearables, Illness, and Recovery,” was a work-in-progress examination of a physical therapy study of patients who use wearable devices to track metrics such as heart rate variability and stress level. Drummond explained how he used Python, R, and scikit-learn to turn the information from reactive to proactive, predicting when one of these patients was becoming sick, had an injury, or was under extended stress. 

Gathering and analyzing data is just one part of the story; being able to communicate that story in a logical way is equally important, and that is something Fairfield professors constantly emphasize, said Cocuzzo.

To learn more about the MS in Business Analytics program, offered both fully online and on-campus, visit fairfield.edu/msba.