IN OPEN ONLINE EDUCATION
Infrastructure for Learning Analytics
Alan Berg
Co chair – SURF SiG LA,
Community Officer Apereo LAI
Program manager LA - UvA
There will be a digital divide
between Higher ED organizations with
data centralism and those without.
Those organizations with data
centralism will have a significant
competitive advantage related to Learning
analytics services.
OU UK
Are we on the wrong side of an emerging digital divide for researchers and
later practitioners?
Before 2018, we aim to become a world leading centre of learning
analytics research, whereby state-of-the-art evidence-based research leads to
advancements of the learning analytics field, transformed into successful
methods, approaches and commercial products. Furthermore, we need to
ensure that research is effectively translated into cost-effective
transformations of the core OU business programme. By developing micro-
level experiments in the Jennie Lee Research Laboratories and meso-
level interventions within the core business model (e.g., Student Experience
Project PVC LT) together with the two other research programmes within IET
(Innovative Pedagogy, Learning in an open world), we aim to become
the leading centre of excellence in learning analytics
as recognised by REF 2020.
•Standards
-xAPI / Caliper
-LTI
•Requirements and hackathons
-Look at all that duplication of functionality and potentially
relearning others lessons
•Retention systems: Open Academic Analytics Initiative
-Predictive Model Markup Language (PMML)
•Dashboards: Open Dashboard
•Broad community and policies– OLA (SoLAR, LACE, JISC,
Apereo LAI, SURF, UvA)
•Learn from formative examples: JISC national Freemium LA
Service
Building large scale systems
The recipe
● Gathers student activity in a secure repository
● Standard protocol
● REST like service for writing and querying
● Can act as the glue for Learning Analytics
Projects
● Good introduction at:
http://tincanapi.com/overview/
What is the experience API
The basic architecture
Learning
Record Store
Mobile
Devices
OtherLMS ALMS X
Dashboard
Application Ethic’s and privacy and
legal
Security
xAPI Benefits
● Decouples activity streams from specific software
(centralizes data)
● Standards approach
● Scalable - But {the LRS market needs to mature}
● Structured data - But {depends on what you send}
● Community
● Clean data for research
● Growing market enthusiasm
● Can work well with LTI dashboards
Learning Tools Interoperability (LTI)
•Learning Tools Interoperability
•Stand alone application that appears inside LMS’s, portal’s etc
•Great for scalability, performance, migration, separation of
concerns
•Large list of compliant tools
•Uses a secret key to secure communication
•LMS passes context information
•Examples: UvAnalytics, Opendashboard
•Homepage: http://www.imsglobal.org/lti/index.html
Context Information
resource_link_id: is a key for the content.
context_id: is the key to the course.
user_id: is the key to the user in the PIWIK database.
roles: Is the set of roles the user has in the course.
lis_person_contact_email_primary: E-mail address of user as defined
in the Blackboard Database.
Full list of variables:
http://www.imsglobal.org/lti/ltiv2p0/uml/purl.imsglobal.org/vocab/lti/v2/variabl
e/index.html
Scalable dashboards using LTI
Learning
Record Store
All Other
enabled
Sakai CLE
Dashboard
Application
LTI
● Scalable
● Lazy coupling
● Has the potential to
work across
applications
IMSGlobal – Caliper
•A competing standard against xAPI, but does
more
•You can find the exact details here:
http://imsglobal.org/caliper/index.html
•Public release very near.
•Recommend keeping track
Hackathons and open dashboard
Requirements count
• Collecting requirements across hackathons
• Sponsored by SURF and UvA
- 2 days. 4 teams. Free to participate
• LAK15 Conference
- 2 days. 4 teams. Conference charge to participate
- http://lak15.solaresearch.org/hackathon
- https://confluence.sakaiproject.org/display/LAI/Learning+Analytics+Initiati
ve
- https://github.com/Apereo-Learning-Analytics-Initiative/
- Open dashboard
- Open LRS
• More hackathons to come
Open Learning Analytics
With learning analytics poised to become a mainstream technology, higher education leaders
from around the world came together following the Learning Analytics and Knowledge (LAK)
2014 conference in Indianapolis, Indiana for an Open Learning Analytics (OLA) Summit.
European OLA summit Nov 2014 in Amsterdam (LACE, Apereo supporting)
• http://solaresearch.org/initiatives/ola/
• In progress
• Heavily influences emerging frameworks such as:
- Apereo Learning Analytics Initiative
- Jisc National Freemium LA infrastructure
JISCNational Framework
QUESTIONS
?
EXTRA
Policy and Infrastructure workshops
http://analytics.jiscinvolve.org/wp/2015/03/03/a-taxonomy-of-ethical-legal-and-
logistical-issues-of-learning-analytics-v1-0/
http://analytics.jiscinvolve.org/wp/2015/02/13/refining-a-systems-architecture-
for-learning-analytics/
JISC Freemium national LA services
Remember
To mention
PMML,
OAAI
ETL and
kettle
http://www.open.ac.uk/students/charter/essential-documents/ethical-use-student-data-learning-analytics-policy
http://analytics.jiscinvolve.org/wp/2015/04/21/code-of-practice-for-learning-analytics-public-consultation/
Context information
oauth_signature : SuSQ5MGM9EwMsfgRpV/y6Eta+NU=
lti_version : LTI-1p0
oauth_signature_method : HMAC-SHA1
tool_consumer_instance_contact_email : blackboard-ic@uva.nl user_id : _129821_1
lti_message_type : basic-lti-launch-request
tool_consumer_instance_guid : 0991091dba98402ab0cd424ecaf4ea89
launch_presentation_document_target : frame
context_title : Learning_Analytics
oauth_version : 1.0
tool_consumer_instance_name : UvA
lis_person_name_full : Fred Peters
resource_link_title : The Learning Game - Gir /test
context_id : _38097_1
roles: urn:lti:role:ims/lis/Instructor
lis_person_contact_email_primary : Made_up_name@uva.nl
lis_person_name_given : Fred
launch_presentation_locale : en-US
oauth_timestamp : 1338360641
Tincan API protocol
http://tincanapi.com/statements-101/
JSON :actor verb object context extensions etc.
{
"actor": {
"name": "Sally Glider",
"mbox": "mailto:sally@example.com"
},
"verb": {
"id": "http://adlnet.gov/expapi/verbs/experienced",
"display": {"en-US": "experienced"}
},
"object": {
"id": "http://example.com/activities/solo-hang-gliding",
"definition": {
"name": { "en-US": "Solo Hang Gliding" }
}
}
}
xAPI + Open Badges
http://learninglocker.net/2014/03/03/tin-badges-or-open-cans-a-technology-tango/
Growing set of
examples

Grand Challenges Learning Analytics

  • 1.
    IN OPEN ONLINEEDUCATION Infrastructure for Learning Analytics Alan Berg Co chair – SURF SiG LA, Community Officer Apereo LAI Program manager LA - UvA
  • 2.
    There will bea digital divide between Higher ED organizations with data centralism and those without. Those organizations with data centralism will have a significant competitive advantage related to Learning analytics services.
  • 3.
    OU UK Are weon the wrong side of an emerging digital divide for researchers and later practitioners? Before 2018, we aim to become a world leading centre of learning analytics research, whereby state-of-the-art evidence-based research leads to advancements of the learning analytics field, transformed into successful methods, approaches and commercial products. Furthermore, we need to ensure that research is effectively translated into cost-effective transformations of the core OU business programme. By developing micro- level experiments in the Jennie Lee Research Laboratories and meso- level interventions within the core business model (e.g., Student Experience Project PVC LT) together with the two other research programmes within IET (Innovative Pedagogy, Learning in an open world), we aim to become the leading centre of excellence in learning analytics as recognised by REF 2020.
  • 4.
    •Standards -xAPI / Caliper -LTI •Requirementsand hackathons -Look at all that duplication of functionality and potentially relearning others lessons •Retention systems: Open Academic Analytics Initiative -Predictive Model Markup Language (PMML) •Dashboards: Open Dashboard •Broad community and policies– OLA (SoLAR, LACE, JISC, Apereo LAI, SURF, UvA) •Learn from formative examples: JISC national Freemium LA Service Building large scale systems The recipe
  • 5.
    ● Gathers studentactivity in a secure repository ● Standard protocol ● REST like service for writing and querying ● Can act as the glue for Learning Analytics Projects ● Good introduction at: http://tincanapi.com/overview/ What is the experience API
  • 6.
    The basic architecture Learning RecordStore Mobile Devices OtherLMS ALMS X Dashboard Application Ethic’s and privacy and legal Security
  • 7.
    xAPI Benefits ● Decouplesactivity streams from specific software (centralizes data) ● Standards approach ● Scalable - But {the LRS market needs to mature} ● Structured data - But {depends on what you send} ● Community ● Clean data for research ● Growing market enthusiasm ● Can work well with LTI dashboards
  • 8.
    Learning Tools Interoperability(LTI) •Learning Tools Interoperability •Stand alone application that appears inside LMS’s, portal’s etc •Great for scalability, performance, migration, separation of concerns •Large list of compliant tools •Uses a secret key to secure communication •LMS passes context information •Examples: UvAnalytics, Opendashboard •Homepage: http://www.imsglobal.org/lti/index.html
  • 9.
    Context Information resource_link_id: isa key for the content. context_id: is the key to the course. user_id: is the key to the user in the PIWIK database. roles: Is the set of roles the user has in the course. lis_person_contact_email_primary: E-mail address of user as defined in the Blackboard Database. Full list of variables: http://www.imsglobal.org/lti/ltiv2p0/uml/purl.imsglobal.org/vocab/lti/v2/variabl e/index.html
  • 10.
    Scalable dashboards usingLTI Learning Record Store All Other enabled Sakai CLE Dashboard Application LTI ● Scalable ● Lazy coupling ● Has the potential to work across applications
  • 11.
    IMSGlobal – Caliper •Acompeting standard against xAPI, but does more •You can find the exact details here: http://imsglobal.org/caliper/index.html •Public release very near. •Recommend keeping track
  • 12.
    Hackathons and opendashboard Requirements count • Collecting requirements across hackathons • Sponsored by SURF and UvA - 2 days. 4 teams. Free to participate • LAK15 Conference - 2 days. 4 teams. Conference charge to participate - http://lak15.solaresearch.org/hackathon - https://confluence.sakaiproject.org/display/LAI/Learning+Analytics+Initiati ve - https://github.com/Apereo-Learning-Analytics-Initiative/ - Open dashboard - Open LRS • More hackathons to come
  • 13.
    Open Learning Analytics Withlearning analytics poised to become a mainstream technology, higher education leaders from around the world came together following the Learning Analytics and Knowledge (LAK) 2014 conference in Indianapolis, Indiana for an Open Learning Analytics (OLA) Summit. European OLA summit Nov 2014 in Amsterdam (LACE, Apereo supporting) • http://solaresearch.org/initiatives/ola/ • In progress • Heavily influences emerging frameworks such as: - Apereo Learning Analytics Initiative - Jisc National Freemium LA infrastructure
  • 14.
  • 15.
  • 16.
  • 17.
    Policy and Infrastructureworkshops http://analytics.jiscinvolve.org/wp/2015/03/03/a-taxonomy-of-ethical-legal-and- logistical-issues-of-learning-analytics-v1-0/ http://analytics.jiscinvolve.org/wp/2015/02/13/refining-a-systems-architecture- for-learning-analytics/
  • 18.
    JISC Freemium nationalLA services Remember To mention PMML, OAAI ETL and kettle
  • 19.
  • 20.
    Context information oauth_signature :SuSQ5MGM9EwMsfgRpV/y6Eta+NU= lti_version : LTI-1p0 oauth_signature_method : HMAC-SHA1 tool_consumer_instance_contact_email : blackboard-ic@uva.nl user_id : _129821_1 lti_message_type : basic-lti-launch-request tool_consumer_instance_guid : 0991091dba98402ab0cd424ecaf4ea89 launch_presentation_document_target : frame context_title : Learning_Analytics oauth_version : 1.0 tool_consumer_instance_name : UvA lis_person_name_full : Fred Peters resource_link_title : The Learning Game - Gir /test context_id : _38097_1 roles: urn:lti:role:ims/lis/Instructor lis_person_contact_email_primary : Made_up_name@uva.nl lis_person_name_given : Fred launch_presentation_locale : en-US oauth_timestamp : 1338360641
  • 21.
    Tincan API protocol http://tincanapi.com/statements-101/ JSON:actor verb object context extensions etc. { "actor": { "name": "Sally Glider", "mbox": "mailto:sally@example.com" }, "verb": { "id": "http://adlnet.gov/expapi/verbs/experienced", "display": {"en-US": "experienced"} }, "object": { "id": "http://example.com/activities/solo-hang-gliding", "definition": { "name": { "en-US": "Solo Hang Gliding" } } } }
  • 22.
    xAPI + OpenBadges http://learninglocker.net/2014/03/03/tin-badges-or-open-cans-a-technology-tango/ Growing set of examples

Editor's Notes

  • #4 The emerging digital divide for Universities. Those with data centralism and those without.
  • #15 Emerging technological solutions are putting pressure on Policy and practices. The JISC infrastructure is forcing changes in policy.
  • #18 Message: Infrastructure and Policy will be closely related
  • #20 Practices and governance are both emerging.
  • #23 Learning Locker Learning Record Store (LRS) http://learninglocker.net/2014/03/03/tin-badges-or-open-cans-a-technology-tango/ “An xAPI statement essentially records an event related to learning. The big idea (I think) is to facilitate the effortless capture of learning events, regardless of learning system or environment. But lonely xAPI statements by themselves are potentially just a sea of meaningless event records . . . . An open badge aims to recognise something, whether it be achievement, attendance or behaviour. But badges gained without authenticated evidence inside are potentially valueless. Together, they provide a meaningful record of achievement, with an authenticated audit trail of activity. The granular statements that xAPI provides support the assessment process of deciding whether someone has achieved a particular badge – or not yet achieved it. Essentially this results in a mapping between the criteria stored in a badge and the activity stored in xAPI statements. This mapping can also be stored inside the badge itself when issued. . . . ADL Design Group We’ve put together a working group for the second round of the ADL’s Experience API design group to do exactly this. Our team, The Experience Badgers, will be working to deliver two key outcomes: 1. To create a prototype for issuing Open Badges based on a pre-specified set of criteria that are evidenced in xAPI statements (i.e. Data comes in, when the criteria is met, the badge is issued to the individual) 2. And to suggest a standard method by which Open Badges are identified in xAPI statements (i.e. what is the appropriate context description and method to describe an Open Badge being issued to a learner, such that 3rd party systems can always ‘detect’ when a statement refers to a badge and potentially always reach the badge assertion).”