Leaping the chasm moving from buzzwords to implementation

Leaping the chasm: moving from buzzwords to implementation of learning analytics George Siemens Technology Enhanced Knowledge Research Institute (TEKRI) Athabasca University February 1, 2012

Slides (with citations and links) http: //www. slideshare. net/gsiemens/educause 2012

1. Roots of learning analytics and context of deployment 2. Becoming at data-intensive university

Won’t make the argument for why analytics are growing

“Imagination no longer comes as cheaply as it did in the past. The slightest move in the virtual landscape has to be paid for in lines of code. ” Latour (2007)

What’s different today? volume (apparently, there’s lots of data) velocity (processing capacity) variety (internet of things, social media) variability (meaning variance)

“Analytics, and the data and research that fuel it, offers the potential to identify broken models and promising practices, to explain them, and to propagate those practices. ” Grajek, 2011

http: //www. dataqualitycampaign. org/ A different way of thinking and functioning

EMC: Data Science Revealed: A Data-Driven Glimpse into the Burgeoning New Field

Reading a book (or any interaction with data) is analytics

Check my activity Predictive Analytics Reporting

Methods, techniques & evidence

Metrics, or analytics on analytics, are hard (and contextual) What is the impact of effective use of data? Argument: “more precise and accurate information should facilitate greater use of information in decision making and therefore lead to higher firm performance. ” Brynjolfsson, Hitt, Kim (2011)

LA resources, publications, archive:

Student success/completion Astin (1996) Tinto (1993)

Distributed, multi-level analytics Suthers & Rosen (2011)

Attention metadata Duval (2011)

Learning networks, crowds, communities Haythornthwaite (2011)

Discourse analysis (automated and manual) De Liddo & Buckingham Shum (2011)

Social learning analytics Buckingham Shum & Ferguson (2011)

Participatory learning and reputation Clow & Makriyannis (2011)

Early warning Macfayden & Dawson (2010) Campbell et al (2006)

Semantic Web to Social Machines “People do the creative work and the machine does the administration” Web=unlimited scaling of info Web should=unlimited social interaction Hendler & Berners-Lee (2010)

1. Roots of learning analytics and context of deployment 2. Becoming at data-intensive university

We collect enough data. We need to focus on connecting. Multiple data sources: Social media University help resources LMS Student information system Course progression, etc

Privacy as a transactional entity Share my data to improve learning support from the university (school)

“All-embracing technique is in fact the consciousness of the mechanized world. Technique integrates everything. It avoids shock and sensational events” Ellul, 1964

Analytics as a complex system: multiple interacting entities, more meaningful when connected

Challenges: Broadening scope of data capture - data outside of the current model of LMS - sociometer: Choudhury & Pentland (2002) - classroom/library/support services, - quantified self Timeliness of data (real-time analytics)

Three communities that don’t communicate Systems/enterprise level Researchers Educators (cobbling)

What does a data-intensive university look like?

Kron, et al (2011)

Acquisition: how do we get the data – structured and unstructured? Storage: how do we store large quantities? Cleaning: how do we get the data in a working format Integration: How do we “harmonize” varying data sets together Analysis: which tools and methods should be used? Representation/visualization: tools and methods to communicate important ideas

“A university where staff and students understand data and, regardless of its volume and diversity, can use it and reuse it, store and curate it, apply and develop the analytical tools to interpret it. ”

43

Principles of a systems-wide analytics tool 1. Algorithms should be open, customizable for context 2. Students should see what the organization sees 3. Analytics engine as a platform: open for all researchers and organizations to build on 4. Connect analytics strategies and tools: APIs 5. Integrate with existing open tools 6. Modularized and extensible

Learning Analytics & Knowledge 2012: Vancouver http: //lak 12. sites. olt. ubc. ca/ Open online course: http: //lak 12. mooc. ca/

Twitter/randomly popular social media: gsiemens www. learninganalytics. net http: //www. solaresearch. org/