The emerging MOOC data/analytics ecosystem

(With thanks to George Siemens, John Stamper and Emily Schneider for helping my thinking on this with their postings to the Learning Analytics group, and yesterday’s conversation at BETT13 with Daphne Koller)

We are about to see the emergence of a MOOC data/analytics ecosystem. Part of the value proposition to partners who sign up to deliver courses on a given MOOC platform is the access to high quality data on what happens with those courses. Arguably, what may differentiate one platform from another is the quality of data, and the provision of integrated analytics tools to help partners make sense of the data, and build their institutional capacity in this field.

(Disclaimer: I’m writing this thought piece not as a employee, but as an academic member of the Open University research team supporting the effort, so these views are my own reflections, from my engagement with the wider MOOC and analytics community)

This piece is not about interoperability standards and protocols, but about how trusted partnerships may emerge to assist ethical data sharing within and between MOOCs, in order to turbocharge educational innovation — one of the primary reasons that institutions are dipping their toes in the MOOC space.

Let’s start with the MOOC-Partner-Course (MPC) framework.

Firstly, in any sector, taking on the burden of hosting a platform means you have privileged access to all the data in it — you have to manage it on a sustainable basis, and there is potential business value in it.

MPC-Framework - MOOC+Partners

Within a MOOC, Partners have access to their own course data:

MPC-Framework - MOOC+Partners+Courses

Now let’s view it in terms other than siloed organisations.

Disciplinary Partnerships: How can we teach X better?

MOOC1-Partner1-Course1 (M1P1C1),  M2P3C7 and M3P27C99 are all teaching aspects of climate change, and agree that there is mutual benefit in sharing the data from their courses. They’re trying to drive innovation in their faculty after all. I could draw the permissions table following the above convention, but you get the idea — access now crosses the MPC boundaries.

Their faculty and learning analytics teams collaborate to create a dataset. In the process they will have to convert several different data formats from 3 MOOC platforms (but MIF eventually makes this straightforward — see below…)

Pedagogical Partnerships: How can we use this learning design better?

M4P2C8 and M8P77C5 are very interested in how best to scaffold working practitioners’ action research, enquiry-based learning in a MOOC context, regardless of discipline. They put out a call for others who share this interest, recruit 6 other MPCs, and design their project.

Other Research Partnerships: RQ-driven collaboration

MOOCs provide an unprecedented opportunity for educational research, and there are many more researchers out there than there are Partners signed up with MOOCs. Researcher1 approaches the faculty behind M4P2C8 and M8P77C5 to see if they are interested in investigating her specific Research Question. They are, but also suggest other MPCs get involved as well, and 3 other MPCs accept the invitation. Partner institutions fund the pilot work, which is so promising it provides the basis for a full grant proposal to a national research council.

Access to the dataset from these or any other researchers in the future obviously requires applications from trained researchers who follow ethical standards.

Analytics Partnerships: How can we optimize the above processes?

As data sharing practice begins to establish itself into a set of patterns, MOOCs 1, 2 and 3 understand that part of their value proposition depends on offering Partners the above flexibility. (Other MOOCs and Partners are pursuing business models for which this makes less sense.) So M1-3 develop the MOOC Interchange Format (MIF) to assist flexible collaborations.

Moreover, M1+2 are particularly interested in embedding integrated, easy to use analytics tools into their Partner admin interfaces, while others leave it to Partners to figure out how to analyse the data with third party tools. The wider learning analytics research and vendor community may in any case have more capacity to develop that third party ecosystem, in which it is expected that tools can import and export MIF.


Does this vision of the emerging ecosystem make sense to you?

What other scenarios do you see?

What do we need to do now in order to make this happen more effectively? (ensuring that we don’t in the process straitjacket MOOCs at this nascent stage in their young lives?)

Comments welcome here
or continue the thread in the Learning Analytics Google Group

2 Responses to “The emerging MOOC data/analytics ecosystem”

  1. The recent AIED2013 MOOC Workshop has two papers from MIT proposing technical standards for extracting and analysing cross-platform MOOC data. Veeramachaneni et al. in particular emphasise the sharing of scripts to help make sense of standardised MOOC data, not the sharing of the raw data itself.

    Kalyan Veeramachaneni, Zachary Pardos, Una-May O’Reilly, Franck Dernoncourt and Colin Taylor (2013). Developing Data Standards and Systems for MOOC Data Science. Proc. 1st Workshop on Massive Open Online Courses at the 16th Annual Conference on Artificial Intelligence in Education (2013). Memphis, TN.

    Piotr Mitros (2013). Insights: An Open Framework for Educational Data Research. Invited talk: 1st Workshop on Massive Open Online Courses at the 16th Annual Conference on Artificial Intelligence in Education (2013). Memphis, TN.

    I’m still figuring out the relationship between these two MIT streams of activity. Sounds like edX Insights has a broader ambition, but that “MOOCdb” should be seen as a candidate for the “missing schema data model” that Piotr appeals for.

  2. “The Maturing of the MOOC” – UK govnt report coinciding with launch, highlighting that the huge opportunity for Learning Analytics

Leave a Reply

You can use these XHTML tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>