Academic advisors play a critical role in guiding students down a fulfilling scholastic path toward graduation and a rewarding career.
At Indiana University, Bloomington, University Division (UD) advisors mentor undergraduates through their first four semesters, offering guidance about balancing classes, schedules, and extracurricular activities within a new social and educational environment. With the assistance of machine-learning algorithms and IU’s supercomputer Carbonate, they have adopted a data-driven approach.
Until recently, University Division advisors bore the sole responsibility for predicting student success within courses, and courses of study. Now, they pair their knowledge of a student’s goals, performance, and risk factors with predictive data to help students make informed choices.
IU staff from the Office of the Vice Provost for Undergraduate Education: Stefano Fiorini Ph.D. (left), Pallavi Chauhan (center), and Adrienne Sewell (right)
When students make informed, data-based decisions with the help of academic advisors, students are empowered to do their best work, making them more likely to succeed at IU.
Dennis Groth, vice provost for undergraduate education at IU
The team at Bloomington Assessment and Research (BAR) within the Office of the Vice Provost for Undergraduate Education, which supports the IU community by providing research and analysis to guide data-driven decisions, analyzed student enrollment data from the past ten years, distilled performance data in different courses, and used the data to predict how a student might perform in a particular course.
To do so, BAR experimented with applying machine-learning algorithms, similar to those that prompt product recommendations online, to student records, predicting that student’s performance in a class. The predictions are not meant to bind students to a path; rather, a requirement of the system was that it focus on helping students along their chosen paths rather than guiding them onto new ones.
After extensive testing and assessment, BAR needed to scale the tool for production, which involved running and provisioning results for 11,000 IUB students at key moments during the semester, and then making that information available to advisors during registration.
Support from the Research Analytics team facilitated the implementation of the predictive algorithms on Carbonate, IU’s large memory computing cluster, significantly reducing the run time and prioritizing the time-sensitive processes to meet campus needs.
Given IU’s ever-increasing enrollment, advisors’ time is at a premium; data insights from predictive models can help advisors identify students in need of assistance quickly, and then focus their efforts on in-person, one-on-one advising. In this way, using Carbonate has helped BAR support advisors and, most importantly, UD students, by advancing emerging computational intensive analytics that support Student Success at IUB.
According to Dennis Groth, vice provost for undergraduate education at IU, “When students make informed, data-based decisions with the help of academic advisors, students are empowered to do their best work, making them more likely to succeed at IU.”