On May 11, 2023, M.S. Data Science students Brittany City and Shara Taylor and M.S. Biomedical Data Science student Wajehah Sanders presented their capstone projects. The event was the first Capstone Presentation for the School of Applied Computational Sciences.
Reaching the presentation event is a major milestone for master’s students in the school.
“As the final course in our master’s programs, the capstone project is the culmination of two years of the academic experience,” says Ashutosh Singhal, chief data officer and instructor for the capstone course. “It emulates a real-world, industry project and each student pursues a topic of their interest.”
In her project, “A Technology Career Predictive Model Using Education, Skills and Big 5 Personality,” City sought to identify the personalities, skills, and interests of technology professionals. Her second goal was to develop a predictive model to suggest viable technology career decisions. She surveyed technology professionals and used Logistic Regression Classification, Random Forest Classification, and Decision Tree models to identify several identifying characteristics. Cosine Similarity was able to recommend a job category with a score of up to 90 percent.
Sanders explored factors related to perinatal depression in her study “Predictors of Maternal Mental Disease: An Analysis of Perinatal Depression and Anxiety.” She performed an observational study using data from 2018 to 2022 from the National Institutes of Health All of Us research program. She used univariate and multivariate logistic regression models. The multivariate analysis found that Black women, when compared to white women, are more likely to have depression with their pregnancy. It also showed statistically significant odds of pregnant women developing depression if they have one of the following conditions: asthma, diabetes, hyperlipidemia, Vitamin D Def, overweight, anxiety, bipolar, depression, and nicotine dependency.
In “Mic Check: The Evolution of Lyricism in Hip-Hop,” Taylor sought to find the words that have been used the most frequently in hip-hop songs over the past five decades. She also assessed whether there was any statistically significant difference in word usage. Her approach included data wrangling, data analysis and hypotheses testing. She collected a dataset from genius.com, that included 1,740,112 words pulled from 10,136 songs by 122 artists.
Taylor found that the top five word frequencies were love, life, money, good and big. She also found significant differences between the 2000s and the 1990s, 2010s and 2020s. Those findings reject the null hypotheses that word frequency is the same for each decade.
Each presentation was includedby a brief question and answer session. Faculty, students and industry partners asked follow up questions and congratulated each presenter on their work.