The Australian Catholic University takes pride in the fact its 35,000 students are spread around seven campuses with a more small-town vibe than its larger rivals, but that source of pride has also contributed to a disparate and complicated data landscape.
Over the last two years, ACU has been working to overcome a lack of enterprise-level integration and data warehousing to establish a data analytics capability, not just to enhance the quality of its forecasting and planning, but to ensure a quality environment for students.
Director at ACU’s Office of Planning and Strategic Management, Dr Omer Yezdani, told iTnews that before overhauling its communications environment and deploying Microsoft’s Power BI, analytics on a variety of metrics could take up to three years to come through.
Part of that, Yezdani said, was simply due to the nature of degrees - they’re not something that will produce results overnight.
However, even five years ago, ACU’s business intelligence capability “was essentially counting rows and columns on a spreadsheet and large printed folders, which would arrive on people's desks in the morning,” he said.
The new approach to data is centred on collecting granular data at every stage of the student lifecycle - from recruitment through to graduation and life as an alumnus - with actionable insights now available in as little as a day.
So far that’s enabled ACU to better predict trends in the popularity of different careers and the degrees behind it - giving the university time to design appropriate recruitment strategies to attract potential students and invest in the teaching staff and campus space to cater to that demand.
And rather than wait until the end of a cohort’s time at the university to obtain a comprehensive picture of the forces behind attrition rates, ACU now gets an updated picture of what engages or deters students on a regular basis throughout the entire degree.
However, internal factors aren’t the only ones that impact on enrolment and retention.
Competition from other universities, student mindsets and the state of the economy all need to be factored into predictions of course popularity.
“And you know, one year, a particular discipline might be a real humdinger, and people are really interested in IT, or law, or criminology or something at a particular university. And then another year, not so much.
“If we have a narrative, for example, where everyone's criticising teachers and saying teachers are no good, they’re not well trained enough, and all that sort of thing, we actually see a decline in enrollments in teacher education.
“So we can't completely account for some of those uncontrolled factors, but what we can do is incorporate predictive analytics that understand some of those patterns over time, because we do tend to see similar cycles if you look with a long enough time horizon.”
Using those analytics to enhance planning and management has enabled the university to realise an ambitious growth strategy, doubling its enrolment figures in the last six years alone, while also supporting the opening of an eighth campus in Rome in 2015.
When it came to choosing a solution to deliver those insights, Yezdani said that Power BI was chosen to complement ACU’s existing use of Microsoft products while also supporting the shift to the cloud.
“We wanted to leverage greater potential into Azure as we moved into the cloud at a similar time, and were looking at the product suite in Azure to use more applications there, including Machine Learning Studio, among other things.
“So it was a neat fit in that respect, cost-effective, reasonably simple - the philosophy behind the product to be able to be used by people who are not data scientists.”
That simplicity resulted in “less force” being used to overhaul of the university’s data capabilities, both from a management perspective and for teaching staff.
Less force, however, meant more time, which Yezdani spent “traversing the university” to ensure he knew the needs of different faculties across separate campuses before deploying anything, and making sure there was a solid foundation of support among users.
“I'd say the groundswell of interest and access to data has probably been more impactful, actually, then just having a single champion to oversee it all.”