I found the “Enhancing Teaching and Learning Through Educational Data Mining and Learning Analytics” (Draft report US Department of Education, 2012) publication an interesting explanation and introduction to ‘educational data mining’ and ‘learning analytics’.

My simple interpretation, is that Data Mining advocates seek to automate learning through extensive analysis of previous learners behaviours captured in learning systems. Whereas advocates of Learning analytics emphasise not prediction but on providing information (often visual) to inform stakeholders for decision-making purposes.

This report seeks to combine these two traditions in the concept of an ‘adaptive learning system’, the model below being self explanatory and at first glance convincing, but I think too tidy in practical terms.

Three deficiencies that stand out for me are:
1. the limited view of learning as shown by the exposure of a ‘Student’ to ‘Content’ as generating ‘student learning data’ – learning as process?
2. the narrow distinction of the learning system – courses (instructional) designers (not the same as administrators or teachers) and the curriculum as expressed by learning outcomes/aims, professional body requirements, etc. which are a significant part of the system;
3. and the fact that online courses seem to be the focus – there are lots of data to be analysed on f2f courses.

I think that the letter point is a more general observation of work on learning analytics.

As someone new to this domain of knowledge, two organisations appear to represent the different traditions, see the Society for Educational Data Mining and the Society for Learning Analytics Research – SoLAR. In a previous post, I conflated both of these under the Analytics badge, but now see more clearly a difference.