University of Kentucky Uses Analytics to Study Student Success Factors
Higher education is beginning to practice what it preaches -- teaching and using analytics for its own benefit. Here's how one university ventured in this new direction.
- By Brian J. Dooley
- April 10, 2018
Universities have a different relationship to analytics than ordinary businesses do because their concerns are different. In addition to internal business and financial issues, universities are responsible for teaching analytics to students and prospective future business leaders. This is particularly important as companies face an analytics skills shortage.
The University of Kentucky has been an innovator in incorporating analytics into both its classrooms and its administration, using SAP's HANA in-memory data management and analytics platform. For the past several years, the university has been experimenting with HANA to provide real-time insight into student behavior and success.
The University's Analytics Mandate
"About four or five years ago, we saw an opportunity to start trying to understand the university a bit better," says Adam Recktenwald, chief technology innovation officer and exec director of enterprise applications. "Our particular interest is in using decision support to aid student success.
"The University of Kentucky is a land-grant institution in higher education. It was founded about 150 years ago with the mission not only to be a top-notch university but also to serve the Commonwealth. We take the responsibility of serving our students seriously. One of the ways we want to help is with the retention of students. Students drop out of college all the time for a variety of reasons, and we wanted to understand why this was and help change their behavior so they would persist and get a college degree -- which is far more valuable for long-term earning potential than a high school diploma."
The university installed an SAP HANA data management platform and analytics solution. The main impetus for changing from earlier Oracle and SQL Server solutions was the extremely high speed of HANA's operation.
"Analytics went from complex overnight routines to real time," says Recktenwald. "Once everything is up and configured, you can make decisions about things with the flip of a switch. It really changed our capability to write analytics applications, allowing us to ask many more questions than are possible with conventional platforms.
"We started to understand the university a lot better, and to run analytics that determine such things as what is keeping students from progressing from one class to another, in moving from one major to another, or the impediments to progression from sophomore to junior to senior. We were able to start digging into all of the things that could influence this, such as when courses are offered, whether or not the campus has an impact, socioeconomic background, and self-test scores."
In-Memory Key to Results
The key to HANA's analytics speed is that it produces results in very fast memory rather than through disk storage. This allows the University of Kentucky to run many more iterations of models and check results quickly. It enables analysts to try many possibilities to understand the factors that lead to either dropping out or graduation.
"Running models again and again makes it possible to improve them," Recktenwald points out. "We have some Oracle queries that took 26 to 28 minutes to run. With this amount of time, there are a limited number of iterations that can be tried during the day as you run the model, check it, and run it again. HANA does this in a few seconds so you can tweak models again and again until you get it right. It's very agile and this is a big advantage."
The University of Kentucky was an early adopter, on the bleeding edge of technology, and although there were challenges using an early system, the benefits were certainly worthwhile. HANA made a great difference for most queries. However, Recktenwald notes that there are still cases which require a standard database approach. He uses the analogy of driving a racecar. The racecar is designed to go fast and does that job well, but it's not really designed to provide safety airbags, air conditioning, or roll-up windows.
"We had an existing Oracle data warehouse and we also had an existing SAP data warehouse environment. We also use Microsoft SQL Server. The Oracle system was retired, but SQL Server is still used to handle certain queries. Each data warehouse and analytics platform has different strengths; we are expanding use of the HANA system, but there is still a place for basic data warehouse operations."
Over the next 6 to 12 months, the university is optimizing its ERP platform to run on HANA and squeeze all of the performance possible from the platform.
Introducing the Student Activity Hub
In the most recent offshoot of SAP's data analytics in the education area, the company has just launched the SAP Student Activity Hub, a new iteration of the platform pioneered by the University of Kentucky.
The solution captures data from sources such as student information systems and learning management systems that may be scattered across a campus. It secures data around individual students' experiences and provides opportunities to build personal profiles of every student from scattered data sources. The Student Activity Hub also integrates well with SAP data warehouses and other tools to provide a wide range of added capabilities such as predictive analytics, artificial intelligence, and machine learning.
Brian J. Dooley is an author, analyst, and journalist with more than 30 years' experience in analyzing and writing about trends in IT. He has written six books, numerous user manuals, hundreds of reports, and more than 1,000 magazine features. You can contact the author at [email protected].