News Story

OU Analyse the UK’s only significant headway in Predictive Analytics – says Higher Ed Commission

Jaisha Bruce, Thursday 28 Jan 2016

The Higher Education Commission is an independent body made up of leaders from the education sector, the business community and the major political parties

On the 26th January the Commission published their fourth report “From Bricks to Clicks – The Potential of Data and Analytics in Higher Education” The publication event was held at the House of Lords and KMi Director, professor John Domingue was there.

“This significant report on the value of data analytics in education is something that will help shape the way universities support students” said professor Domingue “and I am proud that the OU is seen as a leading light in this domain.”

The report offers 12 Recommendations for the HE sector on how to exploit big data in education and presents in detail two case studies: the Student Dashboard implemented at the Nottingham Trent University and the OU Analyse predictive analytics project developed and deployed by the OU’s Knowledge Media Institute.

The report acknowledges that whilst Learning Analytics is in relative infancy it is a powerful way for Universities to achieve their strategic goals, as well as provide real benefits for students and recommends that all HEIs should consider appropriate systems to improve student support and performance.

The Commission points out that the wealth of data generated and potentially processed in education will significantly change the landscape of the HE sector. If properly handled, the benefactors will be both the HE institutions and individual students. In the conclusion, the report states

“For example, predictive analytics can identify which students may not complete their degree on time or even hand in individual assignments, which is already being seen in the UK through the OU Analyse tool. Apart from the OU the Commission does not believe that any UK institution has made significant headway in this area.”

OU Analyse utilises machine-learning based methods for early identification of students at risk of failing. Student demographic data and fine-grained dynamic data taken from their VLE activity are modeled against data from previous students. This is used to build and validate predictive models, which are then applied to the current student cohort, identifying students at risk of failure on a weekly basis and prompting appropriate tutor interventions.