Skip to content

Open Learning Analytics: A proposal

Learning analytics are increasingly relevant, and prominent, in education. Startups and established software vendors are targeting learning analytics in their product offerings for the education and training and development sector. Many of the companies that serve the higher education market– including Sungard, Blackboard, and Pearson – are already heavily committed to analytics. Analytics is quickly becoming a term that gets slapped onto any existing product (remember social media from a few years ago? Suddenly, everything was “social”. In education, everything is becoming “data” and “analytics”). Fortunately, analytics can be much more than a software marketing value-added term.

To some degree, all educators are involved in analytics. It might be as basic as being the end-user of a recommender system when buying a book online or as complex as using curriculum pathways to determine the prospect of student success. And, when we’re not the ones analyzing data, our digital data trails are fodder for others who are. I don’t think I’m over speaking when I state that in five years, analytics tools and suites will be as central in higher education learning management and enterprise resource management systems are today. The reason is simple: learning analytics provide insight into what is happening in the learning process and how teaching and learning strategies impact learner success. Of course, analytics in education go beyond only learner success and can provide insight into the outcome of systemic reform initiatives as well as general resource allocation (I have a short post on my learning analytics site on the topic).

Given the importance of analytics in education, we need to have an early discussion on openness. Why not start with an open system rather than adding openness on as an afterthought once systems are already established? Why not learn from the experiences of previous system-wide software development processes (i.e. LMS) and apply those lessons right up front in the planning process?

To address the need for openness of platforms, algorithms and ensure that the learning process remains a key focus, a group of us have proposed the development of an open learning analytics architecture/platform. We’ve posted our (beta) vision online: Open Learning Analytics: an integrated & modularized platform (.pdf). We are interested in hearing from, and partnering with, others – researchers, educators, universities, schools, startups, and corporate partners (learning and development departments). We have submitted several grant applications and have a few more that will be submitted in the next six months (one early response chastised us for being “too ambitious”. I solidly reject that assertion. Why is it that corporate entities can have ambitious plans but researchers are expected to think in isolated minutiae? Researchers need to think in systems and platforms in order to have an impact).

Additionally, we (Simon Buckingham Shum, Shane Dawson, Erik Duval, Dragan Gasevic, and myself) are offering an open online course on learning analytics starting January 2012. Sign up is available here. Finally, if you’re interested, we are hosting our 2nd conference on Learning Analytics and Knowledge in Vancouver at the end of April.

I’ve embedded a presentation below I did at IADIS conference at East China Normal University in Shanghai this past week on Envisioning a System Wide Learning Analytics Platform (slide 31 details some of the principles behind the project):


  1. Thanks George,

    I think we all need to check and qualify common statements like the following,

    “The reason is simple: learning analytics provide insight into what is happening in the learning process and how teaching and learning strategies impact learner success.”

    The learning process referred to there, is the one designed by educational institutions, the target and the reasoning for bringing data (not evidence) driven analysis on to the behaviors exhibited in those situations.

    I know you’re writing here, is for that particular audience, but I fear that they do not check and qualify their thinking on what learning is, and where it happens, slipping into catch-all terminology, and therefore thinking, that blinds them to the opportunities that have always taken place outside their point if view.

    Four example, your other work in large scale, networked and informal learning collaborations that some call MOOCs (Massive Open Online Courses), or the even more informal learning right across networks and communities, whether online our not. The Open Education Resources University, despite it’s concerning alliances, seems to be arranging toward recognizing these informal approaches to learning, with a view to awarding the papers that certify people. What I how they will want, are tools for people to easily gather their evidence, and present to such accrediting processes, without necessarily submitting to an institutionally designed method if learning.

    I know you’ll agree that a data based perspective in learning would be better if it enabled people to reflect on their own approaches to learning, rather than institutions using it to modify the designs and systems they would have people go through. Hopefully your call here for open data will help facilitate this.

    Given the power those institutions have over many of the material rewards of learning, I think we need to be extra careful to check and qualify what we mean by learning, so as to not see this meme that you are working on, go the same way as other concepts, like ‘social’, or ‘ learner centered’ or ‘ life long learning’, or ‘ personalized learning’, where they have all been absorbed into the power those institutions have.

    Saturday, December 10, 2011 at 3:25 pm | Permalink
  2. Apologies for the typos above, I’m on a phone with predictive text.

    Saturday, December 10, 2011 at 3:33 pm | Permalink
  3. Lisa wrote:

    I agree that analytics can be a powerful tool used especially in university platforms such as Blackboard and ecollege. And I agree with Leigh that we need analytics that provide information as feedback to students so that they may make their own learning adjustments. I liken this process to the use of thinking styles in the sense of Robert Sternberg’s thinking styles. Using his version of thinking styles allows the learners to become aware of their differing strenghts in any given “thinking” situation or activity and note their weaknesses so that they may work toward improvement in thses particular thinking areas/concepts.

    Whatever we educators can do to help our students to learn, especially in their own way (constructivist here), then the better I say. The learning becomes more transformative and inherent long-term.

    Sunday, December 18, 2011 at 4:45 am | Permalink