www.triballabs.net


Tribal Learning Analytics R&D Project
Chris Ballard – Innovation Consultant (Analytics)
6th December 2012




                                                       @chrisaballard
Who are we?




Tribal Learning Analytics R&D Project
Our work with Learning Analytics




Tribal Learning Analytics R&D Project   3
“Every ….. days we create as much
 information as we did from the dawn of
 civilization up until 2003. That’s
 something like five exabytes of data.”


Eric Schmidt (Google CEO)
Do we have Big Data in Higher
Education?




Tribal Learning Analytics R&D Project
Do we have Big Data in Higher
Education?

                   Yes, but…

                   Big is relative.

Tribal Learning Analytics R&D Project
Factors affecting Retention and
Success

                         Academic Integration                        Engagement                    Circumstances




                                                Social Integration                Preparation for HE




Tribal Learning Analytics R&D Project                                                                              7
Factors affecting Retention and
Success

                         Academic Integration                        Engagement                    Circumstances

                                Grades                           VLE Activity                Social Background
                               Progress                        Library Activity                  Proximity
                                                                                               Student Debt




                                                Social Integration                Preparation for HE

                                          Forum interaction                       Demographics
                                           Social networks                        Qualifications
Tribal Learning Analytics R&D Project                                                                              8
Objectives for project
                                                         Supporting the student


                                        Predict which students who may require additional support


                                                          Comparison to peers


                                                     Identify potential problem areas


                                                         Give staff better insight


                                                       Enable “actionable insights”


                                                    Steer students towards success

Tribal Learning Analytics R&D Project                                                               9
Student “Success”
                         Withdrawal


                   True                 False




                         Quantitative



Tribal Learning Analytics R&D Project           10
Student “Success”
                         Withdrawal                                  Success


                   True                 False   Completed   Passed
                                                                         Reached
                                                                         average
                                                                                    Exceeded
                                                                                   expectations




                                                Academic
                                                               Satisfaction            …
                                                 Success



                         Quantitative                            Qualitative



Tribal Learning Analytics R&D Project                                                             11
Quantifying academic success



                                 All students      Cluster       Individual
                                                                  Student


                                                               Attainment of
                                                Median grade
                             Median Grade                          cluster
                                                 for cluster
                                                               median grade




Tribal Learning Analytics R&D Project                                          12
Student Information System                                   Activity Data              Engagement


              Academic                                                                             Academic Integration
            performance at              Course Enrolment      Attendance            VLE Usage
               entrance
                                                                                                   Preparation for HE


                                                                                                   Social Integration
                                                              Contact with
           UCAS Application                   Fees                                 Library Usage
                                                            support services
                                                                                                   Future data sources



          Social background               Assessments      Contact with tutors   Campus PC Usage




             Demographics                  Proximity       Social interaction      Door access




                            Open Data

                      IMD                 Spatial




                                                                   Predictive Model
Tribal Learning Analytics R&D Project                                                                             13
Visualising Predictions
 Predictions should help staff make informed decisions
 Predictions from a model are just part of the picture
 Predictions should be combined with staff experience and knowledge
 Predictions should empower staff to ask the right questions




 Predictions are a tool to help staff understand
 where there might be issues and inform
 subsequent discussions


Tribal Learning Analytics R&D Project                                  14
Tribal Learning Analytics R&D Project   15
Tribal Learning Analytics R&D Project   16
Tribal Learning Analytics R&D Project   17
Tribal Learning Analytics R&D Project   18
Tribal Learning Analytics R&D Project   19
Tribal Learning Analytics R&D Project   20
Summary
 Student Success
         Often focused on “academic success”
         Are the current definitions of student success too simplistic?
 Predictive Model
         The model needs to be “transparent”
         Allow practitioners to see where likely issues may lie
         Combining diverse models results in greater predictive accuracy




Tribal Learning Analytics R&D Project                                       21
Summary
 Data Visualisation for Learning Analytics
         Should be focused on providing information to help inform discussions
         Supplement predictions with analytics based on underlying activity data
         Comparison with cohort enables comparative judgements to be made
 Actionable Insights
         Embedding intervention recording, management and workflow
         Feedback loop to understand whether interventions make a difference




Tribal Learning Analytics R&D Project                                               22
Tribal Learning Analytics R&D Project   23
Chris Ballard
       Innovation Consultant, Tribal




        twitter: @chrisaballard
        blog: www.triballabs.net




Tribal Learning Analytics R&D Project   24

Tribal Learning Analytics R&D Project - SoLAR Storm Presentation

  • 1.
    www.triballabs.net Tribal Learning AnalyticsR&D Project Chris Ballard – Innovation Consultant (Analytics) 6th December 2012 @chrisaballard
  • 2.
    Who are we? TribalLearning Analytics R&D Project
  • 3.
    Our work withLearning Analytics Tribal Learning Analytics R&D Project 3
  • 4.
    “Every ….. dayswe create as much information as we did from the dawn of civilization up until 2003. That’s something like five exabytes of data.” Eric Schmidt (Google CEO)
  • 5.
    Do we haveBig Data in Higher Education? Tribal Learning Analytics R&D Project
  • 6.
    Do we haveBig Data in Higher Education? Yes, but… Big is relative. Tribal Learning Analytics R&D Project
  • 7.
    Factors affecting Retentionand Success Academic Integration Engagement Circumstances Social Integration Preparation for HE Tribal Learning Analytics R&D Project 7
  • 8.
    Factors affecting Retentionand Success Academic Integration Engagement Circumstances Grades VLE Activity Social Background Progress Library Activity Proximity Student Debt Social Integration Preparation for HE Forum interaction Demographics Social networks Qualifications Tribal Learning Analytics R&D Project 8
  • 9.
    Objectives for project Supporting the student Predict which students who may require additional support Comparison to peers Identify potential problem areas Give staff better insight Enable “actionable insights” Steer students towards success Tribal Learning Analytics R&D Project 9
  • 10.
    Student “Success” Withdrawal True False Quantitative Tribal Learning Analytics R&D Project 10
  • 11.
    Student “Success” Withdrawal Success True False Completed Passed Reached average Exceeded expectations Academic Satisfaction … Success Quantitative Qualitative Tribal Learning Analytics R&D Project 11
  • 12.
    Quantifying academic success All students Cluster Individual Student Attainment of Median grade Median Grade cluster for cluster median grade Tribal Learning Analytics R&D Project 12
  • 13.
    Student Information System Activity Data Engagement Academic Academic Integration performance at Course Enrolment Attendance VLE Usage entrance Preparation for HE Social Integration Contact with UCAS Application Fees Library Usage support services Future data sources Social background Assessments Contact with tutors Campus PC Usage Demographics Proximity Social interaction Door access Open Data IMD Spatial Predictive Model Tribal Learning Analytics R&D Project 13
  • 14.
    Visualising Predictions  Predictionsshould help staff make informed decisions  Predictions from a model are just part of the picture  Predictions should be combined with staff experience and knowledge  Predictions should empower staff to ask the right questions Predictions are a tool to help staff understand where there might be issues and inform subsequent discussions Tribal Learning Analytics R&D Project 14
  • 15.
  • 16.
  • 17.
  • 18.
  • 19.
  • 20.
  • 21.
    Summary  Student Success  Often focused on “academic success”  Are the current definitions of student success too simplistic?  Predictive Model  The model needs to be “transparent”  Allow practitioners to see where likely issues may lie  Combining diverse models results in greater predictive accuracy Tribal Learning Analytics R&D Project 21
  • 22.
    Summary  Data Visualisationfor Learning Analytics  Should be focused on providing information to help inform discussions  Supplement predictions with analytics based on underlying activity data  Comparison with cohort enables comparative judgements to be made  Actionable Insights  Embedding intervention recording, management and workflow  Feedback loop to understand whether interventions make a difference Tribal Learning Analytics R&D Project 22
  • 23.
  • 24.
    Chris Ballard Innovation Consultant, Tribal twitter: @chrisaballard blog: www.triballabs.net Tribal Learning Analytics R&D Project 24

Editor's Notes

  • #3 5 exabytes – all words ever spoken by human beings throughout the whole of civilization.
  • #6 Question is does HE have what we might call big data? And also can HE institutions extract knowledge from that data which can be used to help students?Yes, but…Big Data is relative – data sets generally smallerExamples… generally going to centre around electronic interactionsStudent Information System (e.g. SITS),VLE activity – Blackboard/Moodle (logs)Other logs of student activityOnline social interactionLots of opportunities and knowledge can be opened up by analysing multiple data sets which are available in FEWhat does this mean to education?Well, in the same way as user's of amazon, students leave a trail of data resulting from their interactions with university services.We can use this data to help us understand how well students are engaging academically and understand which patterns of engagement lead to improvements in the likely academic success.Learning lessons from how amazon uses data can enable an institution to become smarter about how they use the data they have.This is Learning Analytics - using the data we collect about students, to help us support the student, taking into account their particular needs, background and situation.
  • #7 Question is does HE have what we might call big data? And also can HE institutions extract knowledge from that data which can be used to help students?Yes, but…Big Data is relative – data sets generally smallerExamples… generally going to centre around electronic interactionsStudent Information System (e.g. SITS),VLE activity – Blackboard/Moodle (logs)Other logs of student activityOnline social interactionLots of opportunities and knowledge can be opened up by analysing multiple data sets which are available in FEWhat does this mean to education?Well, in the same way as user's of amazon, students leave a trail of data resulting from their interactions with university services.We can use this data to help us understand how well students are engaging academically and understand which patterns of engagement lead to improvements in the likely academic success.Learning lessons from how amazon uses data can enable an institution to become smarter about how they use the data they have.This is Learning Analytics - using the data we collect about students, to help us support the student, taking into account their particular needs, background and situation.