As with all my other live blogs of this event please read in conjunction with my personal aims

SOLAR (Society for Learning Analytics Research) Update
We heared today about SOLAR FLARE (local meetings including UK), SOLAR STORM (collaborative research effort), solarresearch.org an open access journal to support Learning Analytics. There may be some opportunity for a joint event between JISC and SOLAR in October.

Katy Borner Visual Analytics in support of education
Katy is a big name in the business. A three part talk – motivation, theory, practice of visual analytics. She kicks off with monitoring 3D virtual environments in support of spatial, topical and social navigation and teaching optimization. That’s Second Life. VLearn 3D conference was a 7 hour event where the ‘movements’ of the attendees was tracked and analysed (mapping virtual words and their inhabitants) to help redesign. Presents a Needs Driven Workflow Design as a complex slide as an approach to ‘do’ visualization analytics for a client. All you could ever want to know about how to do data visualisations. The PPT Stack should be useful for the JISC Recon piece.

Student Success System: Risk analytics and Data Visualization using ensembles of predictive models
Empowers student Mentors with predictions based on Sakai LMS records so they can make interventions prior to student failure.
Retrospectively looked at A students, B and C as cumulative attendance records which gives a good correlation. Rules; higher than 85%, lower that 55%, 5% from boundaries these categories provide prioritization of intervention. Closes the gap between problem identification and intervention. Organisational capacity to use the analytics and make the interventions is a larger obstacle than the analytics themselves.
Seems simple, easily achievable and may get the ball rolling toward more complex predictions. Has run for a year so they have no data on impact. Shame.
Slides here

Applying AI to Educational Data: Syllabus Quality Analysis
Aims to address affordability of course provision while ensuring educational quality by asking the questions;

What is a good syllabus?
What are metrics for quality?

Suggests analyzing;
Quality of materials, motivation of instructor and student, scheduling and school facilities

By;
Calculate syllabus consistency, validity, reliability and costs using an intelligent tool.

Then check the validity with expected outcomes of student learning and performance.

Seeing what the system thinks you know – visualizing evidence on an open learner model.
Aposdle is an adaptive learning system used in the workplace. It tracks what people are doing with it and gives recommendations as to what they should be doing next – an adaptive system.

Contentious in it’s assumptions that a good course can be broken down into simple metrics and unusual in that it relies on validity and reliability item analysis techniques which in themselves require a very large sample of normalised data to be reliable.

GLAS (Gradients Learning Analytics System) a first look through an analytics system
Problem to address here is that activities generate large amounts of data, need to discover hidden information, reusing visualisations in different contexts and accommodating differences.
Logs the number of actions / events the students create during a course
Most work just before a required face to face intervention (tutorial), least at holiday time (much of this analytics seems to take a lot of effort to deduce the obvious, apportion the obvious to the individual, then address the obvious, maybe am cynical today!)
One could imagine granting access to personal ‘actions / interventions’ to have them analysed and receive anonymised advice including performance based on peer cohorts