Learning Analytics and
Educational Data Mining:
An Overview of Recent Techniques
Chris&na M. Steiner
, Michael D. Kickmeier-‐Rust, & Dietrich Albert
Graz University of Technology, Austria
www.leas-‐box.eu
Outline
•
Intro
•
Overview of Learning AnalyGcs
•
Learning AnalyGcs and Serious Games
•
Drawbacks and Challenges
Introduction
•
EducaGonal assessment
–
Gathering informaGon about a learner relaGve to
specific competencies, learning objecGves etc.
–
SummaGve assessment
–
FormaGve assessment
•
Computer-‐based educaGon
–
Variety of technologies used for educaGonal
purposes
www.leas-‐box.eu
Introduction
•
Learning AnalyGcs (LA) and
EducaGonal Data Mining (EDM)
–
Making sense of learning data
–
AccounGng for educaGonal data from diversity
of systems/tools
„Big data, applied to
education“
(Horizon Report, 2013)
LA Overview
–
Definition,
Key Concepts
•
LA & EDM
–
LA: „Measurement,
collec&on, analysis
and
repor&ng
of
data about learners and their contexts
, for purposes
of
understanding and op&mising learning
and the
environment in which it occurs.“ (SoLAR)
–
EDM: „Discipline, concerned with developing
methods
for exploring
the unique types of
data
that
come from
educa&onal se<ngs
, and using those methods to
be]er
understand students
, and the
se<ngs
in which
they learn in.“
(Int. EDM Society)
(Cha^ et al., 2012)
LA Overview
LA Overview
–
Stakeholders
& Objectives
•
Learners
•
Teachers
•
EducaGonal insGtuGons
•
Administrators, authoriGes
•
Training developers and providers
www.leas-‐box.eu
•
Monitoring and analysis
•
PredicGon and intervenGon
•
Tutoring and mentoring
•
Assessment and feedback
•
AdaptaGon
•
PersonalisaGon and recommendaGon
•
ReflecGon
LA Overview
–
Stakeholders
& Objectives
LA Overview
-
Data
•
EducaGonally relevant!
•
Centralised vs. distributed
•
Challenge of data integraGon
•
Data set
–
Extensive data
www.leas-‐box.eu
•
Indicators
–
DisposiGonal indicators
–
AcGvity and performance indicators
–
Student artefacts
LA Overview
-
Data
LA Overview
- Analytics
•
PredicGon methods
–
Forecast a certain variable from combinaGon of
indicators
•
Structure discovery
–
DetecGng structure in educaGonal data
•
RelaGonship mining
www.leas-‐box.eu
•
Discovery with models
–
Using results of one analyGcs method within
another analysis
•
DisGllaGon of data for human judgement
–
Providing humans access to reports and
visualisaGons of learner data for judgement
•
Discourse analysis
–
Analysing wri]en evidence of learning
acGviGes and online communicaGon
LA Overview
- Analytics
LA Overview
- Visualisations
•
VisualisaGons make LA results acGonable
–
ReporGng back informaGon to learners, teachers,
other stakeholders
–
Overview on large amounts of data
•
VisualisaGon of
–
AcGvity data
–
Inferences drawn – understanding, competencies
•
VisualisaGons are used
www.leas-‐box.eu
LA and Serious Games
•
Serious games as educaGonal tools
•
LA as an approach for supporGng
performance measurement, assessment, and
improvement in and of serious games
•
Purpose of LA in serious games
–
Measuring student success for feedback to
learners and teachers
–
Game evaluaGon
LA and Serious Games
•
Diversity of games as a challenge
for development of LA approaches and tools
•
Generic approach of LA in serious games
–
UGlising and visualising general interacGon data
–
Teacher-‐defined, game-‐specific assessment rules
•
CombinaGon of game trace variables to obtain new
www.leas-‐box.eu
LA and Serious Games
•
On-‐line/stealth assessment
–
ConGnuous assessment of performance and progress
•
Evidence centred design
•
Micro-‐adapGve assessment and intervenGons
–
Use of learner interacGons to feed and update
the student model (competence, moGvaGon...)
–
SupporGng teaching and learning
•
Monitoring and documenGng learning curve
•
Timely and tailored intervenGons and feedback
•
IndividualisaGon of game experience
LA and Serious Games
•
LA in games sGll at the beginning
•
Serious games as part of mulGple learning
tools and acGviGes
www.leas-‐box.eu
Drawbacks & Challenges
•
Research and pracGce gap
•
Need to answer relevant educaGonal quesGons
•
From user tracking to meaningful insight
–
E.g. on competencies and skills
•
Meaningful integraGon of distributed
learning data
•
Need for user friendly tools
Learning Analytics
Toolbox
•
LEA‘S BOX Project (FP7 STREP)
–
Competence-‐centred and theory-‐grounded LA
•
Competence-‐based Knowledge Space Theory
•
Formal Concept Analysis
–
Novel visualisaGon approaches
www.leas-‐box.eu
Learning Analytics
Toolbox
•
TriangulaGng learning data from
mulGple sources
•
Hybrid approach combining bo]om-‐up and
top-‐down procedures
–
Theory-‐driven competence models
–
Data-‐driven sophisGcaGon and refinement
•
VisualisaGons – OLM
–
Competence structures, concept la^ces
Wrap up
•
Further work on LA is needed to...
–
Integrate LA in educaGonal pracGce
–
Bring together data from a diversity of sources
–
Advance LA for serious games
à
Meaningful, new insights on learning
and teaching
www.leas-‐box.eu
References
• Albert D. & Lukas J., “Knowledge spaces: Theories, empirical research, applicaGons”.
Mahwah: Lawrence Erlbaum Associates, 1999.
• Baker, R., & Siemens, G., “EducaGonal data mining and learning analyGcs”. To appear in
Sawyer, K. (Ed.) Cambridge Handbook of the Learning Sciences: 2nd EdiGon, in press.
• Brown, M., “Learning analyGcs: Moving from concept to pracGce”. EDUCAUSE Learning
IniGaGve. July 2013. Retrieved July 7, 2014 from h]p://net.educause.edu/ir/library/pdf/ ELIB1101.pdf
• Cha^, M.A., Dyckhoff, A.L., Schroeder, U., & Thüs, H. “A reference model for learning
analyGcs.” InternaGonal Journal of Technology Enhanced Learning, 5, 318-‐331, 2012.
• Davenport, T.H., Harris, J.G., & Morison, R., “AnalyGcs at work. Smarter decisions. Be]er
results”. Boston: Harward Business School Publishing, 2010.
• De Liddo, A., Buckingham, Shum, S., Quinto, I., Bachler, M., & Cannavacciuolo, L., “Discourse-‐
centric learning analyGcs”. In Proceedings of the InternaGonal Conference on Learning AnalyGcs and Knowledge, Banff, Alberta, 2011.
• Dyckhoff, A.L., Lukarov, V., Muslim, A., Cha^, M.A, & Schroeder, U. “SupporGng acGon
research with learning analyGcs”. In: Proceedings of the InternaGonal Conference on Learning AnalyGcs and Knowledge (220–229). New York: ACM Press, 2013.
References
• Johnson, L., Adams Becker, S., Cummins, M., Estrada, V., Freeman, A., & Ludgate, H., “NMC
Horizon Report: 2013 Higher EducaGon EdiGon”. AusGn, Texas: The New Media ConsorGum, 2013
• Kickmeier-‐Rust, M.D., & Albert, D., “Micro adapGvity: ProtecGng immersion in didacGcally
adapGve digital educaGonal games”. Journal of Computer Assisted Learning, 26, 95-‐105, 2010.
• Serrano-‐Laguna, A., Torrente, J., Moreno-‐Ger, P., & Fernández-‐Manjón, B., “ApplicaGon of
learning analyGcs in educaGonal videogames”. Entertainment CompuGng, 2014.
• Shute, V. J., Ventura, M., Bauer, M. I., Zapata-‐Rivera, D., “Melding the power of serious games
and embedded assessment to monitor and foster learning: Flow and grow”. In U. Ri]erfeld, M. Cody, & P. Vorderer (Eds.), Serious games: Mechanisms and effects (295-‐321). Mahwah, NJ: Routledge, Taylor and Francis, 2009.
• Wille, R., “Formal concept analysis as mathemaGcal theory of concepts and concept
hierarchies”. In B. Ganter , G. Stumme and R. Wille (eds.), Formal Concept Analysis (1-‐34), Berlin: Springer, 2005.
