One of the unavoidable side effects of our connected digital world and the exponential increase in the number and type of digital devices we carry and use on a regular basis, is the ever increasing digital footprint each of us leaves behind whether comprising our search engine activity, social media streams, GPS data from smart phones or bank card transaction history. According to Donovan (2012:no pagination) “As of 2012, a staggering 2.5 quintillion bytes of data are created daily; 90 percent of the data in the world today have been generated in the last two years alone” and there are four features of this ‘Big Data’ which are categorised as follows
Figure 1: IBM (2011) The four big V’s of data Available at: http://www.ibmbigdatahub.com/sites/default/files/infographic_file/4-Vs-of-big-data.jpg
Our digital footprint is highly valuable to many organisations ranging from business to government and has sparked the growth in the science of Data Analytics or Business Intelligence in recent years as organisations look to segment and analyse this huge quantity of data using a range of quantitative and qualitative techniques in order to identify hidden trends within the data or confirm existing hypotheses.
In its simplest form, teaching is a two-way process that iterates around an instruction – feedback loop; therefore, adopting homogenous approaches to delivering learning is unlikely to achieve either the best potential outcome for learners or enable learning to occur in the most efficient manner as every learner is an individual who has different needs, a unique starting point informed by their previous knowledge and experience, and a preferred way of learning.
In every interaction with learners, whether face-to-face or remotely, synchronous or asynchronous, technology-based or not, each iteration of the instruction-feedback loop produces a stream of ‘data’, the feedback, that competent teachers analyse and use to adapt the next iteration by for example correcting a miscomprehension or extending learning by giving a more complex problem to solve. Frequently this process is quite intuitive in classroom-based situations but in technology-based scenarios directing the learning interaction’s future path in this way relies upon the systematic analysis of some underlying data.
The techniques used by commercial users and government to gain insight from ‘Big Data’ are not lost on educational institutions, Clow (2013) notes that when these general techniques are used in educational contexts they form the basis of learning analytics which Fournier et al (2011:3) define as “the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimising learning and the environments in which it occurs.”
With its relatively recent emergence in the last decade or so, the widespread application of these techniques to educational situations is still evolving rapidly; so much so that Clow (2013) notes that distinct sub-branches of the topic or splits have begun to evolve covering educational data mining which addresses the technical challenges and academic analytics, that focuses on macro level data rather than individual cohorts or students which remains the domain of mainstream learning analytics.
Many of the early learning analytics projects focused predominantly on providing information for teaching staff so that they could take necessary, often offline, intervention action in the case of students who were falling behind. One of the most well reported and arguably most successful initiatives was Purdue University’s Course Signals which modelled and predicted likely student outcomes and provided educators an opportunity to give students real-time feedback via email. The initiative applied learning analytics techniques not only on grades, but also statistical characteristics, previous academic records, and learners’ endeavours as measured by interactions with Blackboard. Clow (2013) reports that by encouraging students to seek help earlier, Course Signals appears to have achieved impressive increases in high grades and reductions in failing grades with learners reporting anecdotally that the support skills were transferable to other courses that did not use the system.
Even after the various successes, learning analytics faces a number of important challenges which Ferguson (2012) categorises as
- Building robust links with the areas of pedagogy, cognition and meta-cognition
- Widening the range of datasets available for analysis away from those recording purely academic performance
- Shifting the focus of analytics from institutions to learners
- Establishing ethical guidelines that keep pace with technological advances and the demand for ever more in-depth analysis
Despite the challenges, learning analytics is evolving, even the definition purpose has been updated; Johnson et al (2016:38) now describe it as “an educational application of web analytics aimed at learner profiling, a process of gathering and analysing details of individual student interactions in online learning activities” with the aim of, among other things, promoting inclusion and targeting students at risk of either failing to realise their potential or even complete at all.
In response to one of the challenges noted by Ferguson (2012), there is now a growing focus of learning analytics is on supporting learners as well as measuring outcomes and there are various initiatives that are producing learner dashboards to show progress and integrating them into learning management systems. For example, Johnson et al (2016:17) cite an initiative by Edinburgh University in conjunction with CogBooks which created “an online adaptive learning and course delivery tool… [whose] dashboard informs students of their progress… while faculty can use the data to improve their teaching”
The influence of learning analytics has grown rapidly over the last decade and the speed at which big data and its analysis impact on learning in the future is only likely to gather pace as new educational technologies such as augmented and virtual reality applications enter widespread use and further contribute to the volume and breadth of data available for analysis.
References
Clow, D. 2013. An overview of learning analytics. Teaching in Higher Education. 18(6) pp683–695
Donovan, F. 2012 Big Data: How It’s Captured, Crunched and Used to Make Business Decisions. 26 June. Technopedia [Online] Available from https://www.techopedia.com/2/28642/enterprise/data-centers/big-data-how-its-captured-crunched-and-used-to-make-business-decisions [Accessed 17/06/16]
Ferguson, R. 2012. Learning analytics: drivers, developments and challenges. International Journal of Technology Enhanced Learning, 4(5/6) pp304–317.
Fournier, H., Kop, R. and Sitlia, H. (2011). The value of learning analytics to networked learning on a personal learning environment. Proceedings of the 1st International Conference on Learning Analytics and Knowledge – LAK ’11. [online] Available at: http://nparc.cisti-icist.nrc-cnrc.gc.ca/eng/view/accepted/?id=7eb5062a-701b-480f-9b10-ec96d72c04b3 [Accessed 31 May 2016].
Johnson, L., Adams Becker, S., Cummins, M., Estrada, V., Freeman, A., and Hall, C. 2016. NMC Horizon Report: 2016 Higher Education Edition. Austin, TX: The New Media Consortium
Digital Learning in Practice 
