Advanced analytics summit brings together experts from numerous universities


Data has increasingly become the driver in businesses worldwide, and colleges and universities are no exception.

Last week, Pitt brought together experts from around the country for the Advanced Analytics Summit: Predicting Student Outcomes with Better Analysis.

Stephen Wisniewski, Pitt’s vice provost for Data and Information, chaired the event. He said this is a first for Pitt, and while there have been other gatherings to discuss analytics in higher ed, “I don’t think they were as broad as this.”

Speakers came from as far away as the Arizona State University, the University of Texas and Texas Tech, and as close by as Carnegie Mellon, West Virginia and Ohio State universities. Joseph J. McCarthy, vice provost for Undergraduate Studies, and Marc Harding, vice provost for Enrollment, also were speakers.

The keynote speaker, Jaime Casap, has the lofty title of Global Education Evangelist at Google Inc. Casap preaches about technology and education, and how they can work together.

Casap, who is a first-generation American and was raised in New York City’s Hell’s Kitchen, said that education is the most important thing and can have an impact for generations.

He told the gathering on Oct. 11 that we have long known that education has to be relevant and personal, but now technology is starting to make that possible.

“We waited for people to teach us,” Casap said. “This generation just starts learning (through technology). This is the world they were born into.”

He said Generation Z students now value experiences and programs, such as early internships and social service projects, when looking at colleges. They also look at outcomes — graduation rates, where students go after graduation, etc.

Casap said higher education needs to adapt by:

  • Converting information into intelligence that students can use.
  • Using innovation to bring education to the next level. Will there even be classrooms in the future?
  • Moving beyond the Caucasian model of recruiting, to reach out to other ethnicities where they live culturally.
  • Engaging students as lifelong learners, to keep them connected to their alma mater for continuing education and mentorship.

Stephen WisniewskiWiesniewski said the conference was aimed at learning from each other and looking at best practices, but he also wanted to “showcase what we’re doing here at Pitt. I think we are trying to advance this area in a very targeted fashion.”

Much of the data analysis in higher education has focused on student enrollment and retention, and how to use that data to intervene proactively with students who are at risk for poor academic performance or low institutional engagement. Some programs alert advisers when students are in danger of failing a course or flunking out, or recommend majors to students based on their interests and academic performance. 

Wisniewski said when he came to Pitt a few years ago, he started looking at what was being done in analytics around the country. “People were saying they were doing advanced analytics, but if you talk to someone who does analysis, it wasn’t truly advanced,” he said.

He found there was a desire to do more advanced work, using machine learning and artificial intelligence, but the implementation was struggling.

“I thought wouldn’t it be nice to get people together who are trying to do this to share success stories and failures, so we can learn from each other and work as a field to develop best practices,” Wisniewski said.

The primary focus, Wisniewski said, is on “student life cycle” — “how do we get them here, how do we keep them here and how do we assure they succeed when they’re here and when they leave.” He said this focus ties into the Plan for Pitt, which stresses students “lead lives of impact” after they leave the University.

“We’re trying to help our students through analytics to succeed and be happy,” Wisniewski said.

“We know over the past four or five years what has happened with our student body, and we build a model based on that population,” he said. That model is then applied to existing students. Based on similar characteristics, the data can estimate the probability of a student having a certain type of event.

For instance, if a student has similar characteristics to people who have had problems graduating in four years, then Pitt can look at what issues the student is having to prevent them from graduating on time and what can be done to help fix those.

Currently, most of the data Pitt gathers comes through PeopleSoft. There are other areas that they’d like to look at, but they need to make sure the date is secure, and that the University has the proper access and is complying with the Family Educational Rights and Privacy Act (FERPA).

Susan Jones is editor of the University Times. Reach her at or 412-648-4294.