Last week, I shared a meal with a friend during which something rare happened— we started discussing principal component analysis. Perhaps others frequently converse on this topic, but for me, it was a first. I attributed this occurrence to the fact that we are both students in technical fields, albeit rather different in scope— I am a Master of Science in Analytics (MSA) student at the Institute for Advanced Analytics (IAA), and he is a PhD candidate in Duke’s Computer Science (CS) department. However, this sparked my curiosity to understand more about the differences (and similarities) between the IAA MSA and the CS PhD student experiences.
During our discussion, we came to appreciate the complexities of both our respective academic paths—from the applicability of the IAA practicum’s team-centric model and fast-paced environment to the PhD program’s comprehensive and guided approach that prepares students for both industry and academic positions.
The schedules for IAA students are planned out so that almost all days are spent in person at Centennial Campus. A typical day starts at 9 AM and is packed until ending at 4 PM, but that schedule is flexible depending on guest speakers, exams, and project-related meetings. PhD candidates like my friend may not have any classes depending on what year they are in. They may have to attend a class to serve as a teaching assistant or host office hours either in-person or virtually a few times a week. They then have the flexibility to head to their office space, where they work, mostly individually, on the research project the lab is working on. My friend has a weekly lab meeting where work that has been done is presented. His research is in partnership with the Duke Center for Autism and Brain Development, and his lab has a weekly meeting to present technical team research to the clinical team. This process sounds somewhat similar to biweekly meetings IAA students have with their practicum sponsors, though the end goal is different.
The practicum project is a team-based project that spans almost the entirety of the 10-month program. Teams work with an industry sponsor to solve a business problem using analytic methods.
The practicum project team is assigned by faculty and staff working with Dr. Rappa, the director of the MSA program. However, typically with PhD programs, research intentions are discussed prior to admission, and this research project can span more than five years, depending on the guidance of the principal investigator and other professors and postdoctoral students in the lab group. My friend does work in a team setting with a few other graduate students in the lab several times a week, but other than that, he is entirely on his own to manage his time with relevant research papers, technical work on ongoing projects, and data annotation and labeling for future project improvements. Students at the IAA have similar guidance from faculty and staff, particularly in the early stages of the practicum project with mandated faculty review and biweekly meetings to ensure the work backlog is planned out.
One reason the programs differ is because the intended outcomes are different post-graduation. PhD programs open doors into academia that the MSA is not intended to do. Almost all students from the MSA go straight into industry roles, with a small number ending up in academia. The curriculum in each program supports what students may experience in their postgraduate lives, PhD programs gearing more towards academic prospects such as individual research and publishing papers, and the IAA focusing mostly on data science applications in a business setting.
Both the MSA and PhD programs have significant merits, but depending on the goals of the student, one can be a better fit than the other. Ultimately, in the field of analytics, it is important to learn from one another, especially given how our experiences outside of academics can inform our knowledge of various programming methods.
Columnist: Xueyang Li