Episode 8: IT IQ Series – Predictive analytics: the “cheat-sheet” for more personalised teaching?

Summary: By detecting learning difficulties before it’s too late, predictive analytics can help teachers gain a far deeper understanding of their students–and how to solve their issues.

Can technology help teachers get the best out of their students? With Australia’s education system struggling to compete globally, many teachers face growing pressure to do more, with less, for an ever-growing number of students under their care. That’s especially true for low socioeconomic zones where only 60% of students finish school, and where teachers face the greatest shortage of resources. These teachers, however, have a new and promising tool at their disposal: data analytics.

“Educators already know that factors such as learning styles and absorption rates impact how effectively students end up performing. But it’s the lack of visibility toward complicated socio-economic or geographic factors that frustrates educators,” says Peter Manoukian, an Account Technology Strategist at Microsoft. “Right now, our approach to these issues is purely reactive: we can only intervene when performance starts to slide. But when you apply analytics to school and classroom data, you start to see the warning signs of academic decline before it happens. That gives teachers a valuable window of opportunity to make a change.”

Predictive analytics platforms have already been used to great effect by different industries to understand consumer behaviour and improve decision-making. With the right data, the same could apply to Australia’s classrooms.

“Schools already have a wealth of historical student data that can be immediately fed into a predictive analytics platform to create reports on where slippages and dropouts have occurred in the past, and more importantly where slippage may occur in the future” explains Manoukian. “Educators want an intuitive, easy method to analyse this student data to confirm their daily observations of students, and to use that confirmation to justify taking action to help declining students.”

From Classroom to Community

Educators can only do so, however, if their analytics tools give them clear answers without requiring overly technical know-how. Some analytics tools in the market, like Microsoft Power BI, already offer analytics-made-easy without any coding knowledge required, helping educators dive in and discover ways to help their students as quickly as possible. Teachers will benefit the most from platforms which present their data and insights through dashboards, charts, and other visual forms of reporting that they–and their principals and other superiors–can quickly understand and act upon.

“Educators can point Power BI to existing Excel worksheets for the subject that they teach, type questions into the tool (using natural language), and find any trends or insights within the data,” says Manoukian. “They can also combine data at different levels, from the student to the entire school: that lets them get a bigger picture of where a particular student might fit within the broader trend, then build a lesson plan that will address their weaknesses or improve their strengths in class. But the more complex these data sets get, the more important visual communication becomes to getting everyone, from teachers to principals, on board with the needed changes.”

But what about long-term challenges such as declining grades or high attrition rates? Australian schools that employ predictive analytics would be in a better position to take proactive action, instead of resorting to a passive ‘wait-and-see’ method. That proactive position starts, says Manoukian, with assessing students’ broader circumstances.

“Through our work with schools around Australia, the Azure team has ended up handling a range of data sets including nutrition, socio-economic l and even social data”  Manoukian explains. “When we feed these seemingly unrelated data sets into a platform like Azure Machine Learning, we can predict various potential problem sites with a reasonably high degree of accuracy. That, in turn, helps schools deploy their resources as well as work with other parties, like government agencies, to remove underlying roadblocks to students’ learning.”

That said, schools can’t treat analytics as a ‘fire and forget’ solution. Teachers and community leaders must assess the actual impact of their policies or processes, then use the results as feedback to further improve the analytics models behind their decision-making. In that regard, analytics and students’ results have one thing in common: you only get out what you put in.

“When working with analytics and machine learning, you’ll need to continuously pump in often a wide range of data to train the machine and allow it to make better predictions,” cautions Manoukian. “The earlier Australian schools start, the longer they’ll have to accumulate that data, and the more effective their efforts to improve students’ performance over time. Change may not come overnight, but the results for our most vulnerable and at-risk students will be well worth the investment.”

Watch Peter Manoukian speak further on the role of Education Analytics and how it can help Australian schools  on our YouTube channel.

Get started on Microsoft Power BI, and learn how to start using the tool to help your class on the Microsoft Virtual Academy.To get started on Azure and Azure Machine Learning, by connecting with our locally certified consultants today.

Our mission at Microsoft is to equip and empower educators to shape and assure the success of every student. Any teacher can join our effort with free Office 365 Education, find affordable Windows devices and connect with others on the Educator Community for free training and classroom resources. Follow us on Facebook and Twitter for our latest updates.