Handbook of Educational Data Mining [Romero, Ventura, Pechenizkiy & Baker 2010-10-25]

UNDERSTANDING COMPLEX DATASETS: DATA MINING WITH MATRIX DECOMPOSITIONS David Skillicorn COMPUTATIONAL METHODS OF FEATURE SELECTION Huan Liu and Hiroshi Motoda CONSTRAINED CLUSTERING: A

Educational Data Mining

UNDERSTANDING COMPLEX DATASETS:

DATA MINING WITH MATRIX DECOMPOSITIONS

David Skillicorn

COMPUTATIONAL METHODS OF FEATURE

SELECTION

Huan Liu and Hiroshi Motoda

CONSTRAINED CLUSTERING: ADVANCES IN

ALGORITHMS, THEORY, AND APPLICATIONS

Sugato Basu, Ian Davidson, and Kiri L Wagstaff

KNOWLEDGE DISCOVERY FOR

COUNTERTERRORISM AND LAW ENFORCEMENT

David Skillicorn

MULTIMEDIA DATA MINING: A SYSTEMATIC

INTRODUCTION TO CONCEPTS AND THEORY

Zhongfei Zhang and Ruofei Zhang

NEXT GENERATION OF DATA MINING

Hillol Kargupta, Jiawei Han, Philip S Yu,

Rajeev Motwani, and Vipin Kumar

DATA MINING FOR DESIGN AND MARKETING

Yukio Ohsawa and Katsutoshi Yada

THE TOP TEN ALGORITHMS IN DATA MINING

Xindong Wu and Vipin Kumar

GEOGRAPHIC DATA MINING AND

KNOWLEDGE DISCOVERY, SECOND EDITION

Harvey J Miller and Jiawei Han

TEXT MINING: CLASSIFICATION, CLUSTERING, AND APPLICATIONS

Ashok N Srivastava and Mehran Sahami

BIOLOGICAL DATA MINING

Jake Y Chen and Stefano Lonardi

INFORMATION DISCOVERY ON ELECTRONIC HEALTH RECORDS

Bo Long, Zhongfei Zhang, and Philip S Yu

KNOWLEDGE DISCOVERY FROM DATA STREAMS

HANDBOOK OF EDUCATIONAL DATA MINING

Cristóbal Romero, Sebastian Ventura, Mykola Pechenizkiy, and Ryan S.J.d Baker

PUBLISHED TITLES

SERIES EDITOR

Vipin KumarUniversity of Minnesota Department of Computer Science and Engineering Minneapolis, Minnesota, U.S.A

AIMS AND SCOPE

This series aims to capture new developments and applications in data mining and knowledge discovery, while summarizing the computational tools and techniques useful in data analysis This series encourages the integration of mathematical, statistical, and computational methods and techniques through the publication of a broad range of textbooks, reference works, and hand-books The inclusion of concrete examples and applications is highly encouraged The scope of the series includes, but is not limited to, titles in the areas of data mining and knowledge discovery methods and applications, modeling, algorithms, theory and foundations, data and knowledge visualization, data mining systems and tools, and privacy and security issues

Edited by

Cristóbal Romero, Sebastian Ventura,

Mykola Pechenizkiy, and Ryan S.J.d Baker

Handbook of Educational Data Mining

CRC Press

Taylor & Francis Group

6000 Broken Sound Parkway NW, Suite 300

Boca Raton, FL 33487-2742

© 2011 by Taylor and Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S Government works

Printed in the United States of America on acid-free paper

10 9 8 7 6 5 4 3 2 1

International Standard Book Number: 978-1-4398-0457-5 (Hardback)

This book contains information obtained from authentic and highly regarded sources Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the valid- ity of all materials or the consequences of their use The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint.

Except as permitted under U.S Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or lized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopy- ing, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers.

uti-For permission to photocopy or use material electronically from this work, please access www.copyright.com (http:// www.copyright.com/) or contact the Copyright Clearance Center, Inc (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400 CCC is a not-for-profit organization that provides licenses and registration for a variety of users For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for

identification and explanation without intent to infringe.

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com

To my wife, Inma, and my daughter, Marta

Editors xv

Contributors xvii

1 Introduction 1

Cristóbal Romero, Sebastián Ventura, Mykola Pechenizkiy, and Ryan S J d Baker Part I Basic Techniques, Surveys and Tutorials 2 Visualization in Educational Environments 9

Riccardo Mazza 3 Basics of Statistical Analysis of Interactions Data from Web-Based Learning Environments 27

Judy Sheard 4 A Data Repository for the EDM Community: The PSLC DataShop 43

Kenneth R Koedinger, Ryan S J d Baker, Kyle Cunningham, Alida Skogsholm, Brett Leber, and John Stamper 5 Classifiers for Educational Data Mining 57

Wilhelmiina Hämäläinen and Mikko Vinni 6 Clustering Educational Data 75

Alfredo Vellido, Félix Castro, and Àngela Nebot 7 Association Rule Mining in Learning Management Systems 93

Enrique García, Cristóbal Romero, Sebastián Ventura, Carlos de Castro, and Toon Calders 8 Sequential Pattern Analysis of Learning Logs: Methodology and Applications 107

Mingming Zhou, Yabo Xu, John C Nesbit, and Philip H Winne 9 Process Mining from Educational Data 123

Nikola Trcˇka, Mykola Pechenizkiy, and Wil van der Aalst 10 Modeling Hierarchy and Dependence among Task Responses in Educational Data Mining 143

Brian W Junker

Part II Case Studies

11 Novel Derivation and Application of Skill Matrices: The q-Matrix Method 159

Tiffany Barnes

12 Educational Data Mining to Support Group Work in Software

Development Projects 173

Judy Kay, Irena Koprinska, and Kalina Yacef

13 Multi-Instance Learning versus Single-Instance Learning for Predicting

the Student’s Performance 187

Amelia Zafra, Cristóbal Romero, and Sebastián Ventura

14 A Response-Time Model for Bottom-Out Hints as Worked Examples 201

Benjamin Shih, Kenneth R Koedinger, and Richard Scheines

15 Automatic Recognition of Learner Types in Exploratory Learning

Environments 213

Saleema Amershi and Cristina Conati

16 Modeling Affect by Mining Students’ Interactions within Learning

Agathe Merceron and Kalina Yacef

18 Data Mining for Contextual Educational Recommendation and Evaluation

Strategies 257

Tiffany Y Tang and Gordon G McCalla

19 Link Recommendation in E-Learning Systems Based on Content-Based

Student Profiles 273

Daniela Godoy and Analía Amandi

20 Log-Based Assessment of Motivation in Online Learning 287

Arnon Hershkovitz and Rafi Nachmias

21 Mining Student Discussions for Profiling Participation and Scaffolding

Learning 299

Jihie Kim, Erin Shaw, and Sujith Ravi

22 Analysis of Log Data from a Web-Based Learning Environment:

A Case Study 311

Judy Sheard

23 Bayesian Networks and Linear Regression Models of Students’ Goals,

Moods, and Emotions 323

Ivon Arroyo, David G Cooper, Winslow Burleson, and Beverly P Woolf

24 Capturing and Analyzing Student Behavior in a Virtual Learning

Environment: A Case Study on Usage of Library Resources 339

David Masip, Julià Minguillón, and Enric Mor

25 Anticipating Students’ Failure As Soon As Possible 353

Cláudia Antunes

26 Using Decision Trees for Improving AEH Courses 365

Javier Bravo, César Vialardi, and Alvaro Ortigosa

27 Validation Issues in Educational Data Mining: The Case of HTML-Tutor

and iHelp 377

Mihaela Cocea and Stephan Weibelzahl

28 Lessons from Project LISTEN’s Session Browser 389

Jack Mostow, Joseph E Beck, Andrew Cuneo, Evandro Gouvea, Cecily Heiner, and

Octavio Juarez

29 Using Fine-Grained Skill Models to Fit Student Performance with Bayesian Networks 417

Zachary A Pardos, Neil T Heffernan, Brigham S Anderson, and Cristina L Heffernan

30 Mining for Patterns of Incorrect Response in Diagnostic Assessment Data 427

Tara M Madhyastha and Earl Hunt

31 Machine-Learning Assessment of Students’ Behavior within Interactive

Learning Environments 441

Manolis Mavrikis

32 Learning Procedural Knowledge from User Solutions to Ill-Defined Tasks

in a Simulated Robotic Manipulator 451

Philippe Fournier-Viger, Roger Nkambou, and Engelbert Mephu Nguifo

33 Using Markov Decision Processes for Automatic Hint Generation 467

Tiffany Barnes, John Stamper, and Marvin Croy

34 Data Mining Learning Objects 481

Manuel E Prieto, Alfredo Zapata, and Victor H Menendez

35 An Adaptive Bayesian Student Model for Discovering the Student’s

Learning Style and Preferences 493

Cristina Carmona, Gladys Castillo, and Eva Millán

Index 505

The Purpose of This Book

tional.data.mining.(EDM) The.primary.goal.of.EDM.is.to.use.large-scale.educational.data.sets.to.better.understand.learning.and.to.provide.information.about.the.learning.process Although.researchers.have.been.studying.human.learning.for.over.a.century,.what.is.differ-ent.about.EDM.is.that.it.makes.use.not.of.experimental.subjects.learning.a.contrived.task.for.20.minutes.in.a.lab.setting;.rather,.it.typically.uses.data.from.students.learning.school.sub-jects,.often.over.the.course.of.an.entire.school.year For.example,.it.is.possible.to.observe.stu-dents.learning.a.skill.over.an.eight-month.interval.and.make.discoveries.about.what.types.of.activities.result.in.better.long-term.learning,.to.learn.about.the.impact.of.what.time.students.start.their.homework.has.on.classroom.performance,.or.to.understand.how.the.length.of.time.students.spend.reading.feedback.on.their.work.impacts.the.quality.of.their.later.efforts.In.order.to.conduct.EDM,.researchers.use.a.variety.of.sources.of.data.such.as.intelli-gent.computer.tutors,.classic.computer-based.educational.systems,.online.class.discussion.forums,.electronic.teacher.gradebooks,.school-level.data.on.student.enrollment,.and.stan-dardized.tests Many.of.these.sources.have.existed.for.decades.or,.in.the.case.of.standard-ized.testing,.about.2000.years What.has.recently.changed.is.the.rapid.improvement.in.storage.and.communication.provided.by.computers,.which.greatly.simplifies.the.task.of.collecting.and.collating.large.data.sets This.explosion.of.data.has.revolutionized.the.way.we.study.the.learning.process

The.goal.of.this.book.is.to.provide.an.overview.of.the.current.state.of.knowledge.of.educa-In.many.ways,.this.change.parallels.that.of.bioinformatics.20.years.earlier:.an.explosion.of.available.data.revolutionized.how.much.research.in.biology.was.conducted However,.the.larger.number.of.data.was.only.part.of.the.story It.was.also.necessary.to.discover,.adapt, or invent computational techniques for analyzing and understanding this new,.vast.quantity.of.data Bioinformatics.did.this.by.applying.computer.science.techniques.such.as.data.mining.and.pattern.recognition.to.the.data,.and.the.result.has.revolutionized.research in biology Similarly, EDM has the necessary sources of data More and more.schools.are.using.educational.software.that.is.capable.of.recording.for.later.analysis.every.action.by.the.student.and.the.computer Within.the.United.States,.an.emphasis.on.educa-tional.accountability.and.high.stakes.standardized.tests.has.resulted.in.large.electronic.databases.of.student.performance In.addition.to.these.data,.we.need.the.appropriate.com-putational.and.statistical.frameworks.and.techniques.to.make.sense.of.the.data,.as.well.as.researchers.to.ask.the.right.questions.of.the.data

munity,.as.can.be.seen.by.the.chapter.authors.of.this.book,.is.composed.of.people.from.multiple.disciplines Computer.science.provides.expertise.in.working.with.large.quanti-ties of data, both in terms of machine learning and data-mining techniques that scale.gracefully.to.data.sets.with.millions.of.records,.as.well.as.address.real-world.concerns.such.as.“scrubbing”.data.to.ensure.systematic.errors.in.the.source.data.do.not.lead.to.erro-neous.results Statisticians.and.psychometricians.provide.expertise.in.understanding.how.to.properly.analyze.complex.study.designs,.and.properly.adjust.for.the.fact.that.most.edu-cational.data.are.not.from.a.classic.randomized.controlled.study These.two.communities

are.strong.in.statistical.and.computational.techniques,.but.techniques.and.data.are.not.suf-Main Avenues of Research in Educational Data Mining

There.are.three.major.avenues.of.research.in.EDM They.nicely.align.with.the.classic.who–what–where–when.interrogatives

ing.which.ones.are.best.suited.to.working.with.large.educational.data.sets,.and.finding.best.practices.for.evaluation.metrics.and.model.fitting Examples.of.such.efforts.include.experimenting.with.different.visualization.techniques.for.how.to.look.at.and.make.sense.of.the.data Since.educational.data.sets.are.often.longitudinal,.encompassing.months.and.sometimes.years,.and.rich.interactions.with.the.student.can.occur.during.that.time,.some.means.of.making.sense.of.the.data.is.needed Another.common.approach.in.EDM.is.using.variants.of.learning.curves.to.track.changes.in.student.knowledge Learning.curves.are.some.of.the.oldest.techniques.in.cognitive.psychology,.so.EDM.efforts.focus.on.examin-ing.more.flexible.functional.forms,.and.discovering.what.other.factors,.such.as.student.engagement.with.the.learning.process,.are.important.to.include One.difficulty.with.com-plex.modeling.in.EDM.is.there.is.often.no.way.of.determining.the.best.parameters.for.a.particular.model Well-known.techniques.such.as.hill.climbing.can.become.trapped.in.local.maxima Thus,.empirical.work.about.which.model-fitting.techniques.perform.well.for.EDM.tasks.is.necessary

The.first.avenue.is.work.on.developing.computational.tools.and.techniques,.determin-sary.foundation.to.EDM Work.in.this.area.focuses.on.how.we.can.extract.information.from.data At.present,.although.a.majority.of.EDM.research.is.in.this.avenue,.the.other.two.are.not.less.important—just.less.explored

This.work.on.extending.and.better.understanding.our.computational.toolkit.is.a.neces-The.second.avenue.is.determining.what.questions.we.should.ask.the.data There.are.several obvious candidates: Does the class understand the material well enough to go.on?.Do.any.students.require.remedial.instruction?.Which.students.are.likely.to.need.aca-demic counseling to complete school successfully? These are questions that have been.asked.and.answered.by.teachers.for.millennia EDM.certainly.enables.us.to.be.data.driven.and.to.answer.such.questions.more.accurately;.however,.EDM’s.potential.is.much.greater The.enormous.data.and.computational.resources.are.a.tremendous.opportunity,.and.one.of.the.hardest.tasks.is.capitalizing.on.it:.what.are.new.and.interesting.questions.we.can.answer.by.using.EDM?.For.example,.in.educational.settings.there.are.many.advantages.of.group.projects Drawbacks.are.that.it.can.be.hard.to.attribute.credit.and,.perhaps.more.importantly,.to.determine.which.groups.are.having.difficulties—perhaps.even.before.the.group.itself.realizes A.tool.that.is.able.to.analyze.student.conversations.and.activity,.and.automatically.highlight.potential.problems.for.the.instructor.would.be.powerful,.and.has.no.good.analog.in.the.days.before.computers.and.records.of.past.student.collaborations.were.easily.available Looking.into.the.future,.it.would.be.useful.if.we.could.determine

if a particular.student.would be.better.served.by.having.a different.classroom.teacher,.not.because.one.teacher.is.overall.a.better.choice,.but.because.for.this.type.of.student.the

teacher.is.a.better.choice The.first.example.is.at.the.edge.of.what.EDM.is.capable;.the.sec-This.job.of.expanding.our.horizons.and.determining.what.are.new,.exciting.questions.to.ask.the.data.is.necessary.for.EDM.to.grow

The.third.avenue.of.EDM.is.finding.who.are.educational.stakeholders.that.could.benefit.from.the.richer.reporting.made.possible.with.EDM Obvious.interested.parties.are.stu-dents.and.teachers However,.what.about.the.students’.parents?.Would.it.make.sense.for.them.to.receive.reports?.Aside.from.report.cards.and.parent–teacher.conferences,.there.is.little.communication.to.parents.about.their.child’s.performance Most.parents.are.too.busy.for.a.detailed.report.of.their.child’s.school.day,.but.what.about.some.distilled.infor-mation?.A.system.that.informed.parents.if.their.child.did.not.complete.the.homework.that.was.due.that.day.could.be.beneficial Similarly,.if.a.student’s.performance.notice-ably.declines,.such.a.change.would.be.detectable.using.EDM.and.the.parents.could.be.informed Other.stakeholders.include.school.principals,.who.could.be.informed.of.teach-ers.who.were.struggling.relative.to.peers,.and.areas.in.which.the.school.was.performing.poorly Finally,.there.are.the.students.themselves Although.students.currently.receive.an.array.of.grades.on.homework,.quizzes,.and.exams,.they.receive.much.less.larger-grain.information, such as using the student’s past.performance.to.suggest which classes to.take,.or.that.the.student’s.homework.scores.are.lower.than.expected.based.on.exam.per-formance Note.that.such.features.also.change.the.context.of.educational.data.from.some-thing.that.is.used.in.the.classroom,.to.something.that.is.potentially.used.in.a.completely.different.place

Research.in.this.area.focuses.on.expanding.the.list.of.stakeholders.for.whom.we.can.provide information, and where this information is received Although there is much.potential.work.in.this.area.that.is.not.technically.demanding,.notifying.parents.of.missed.homework.assignments.is.simple.enough,.such.work.has.to.integrate.with.a.school’s.IT.infrastructure,.and.changes.the.ground.rules Previously,.teachers.and.students.controlled.information.flow.to.parents;.now.parents.are.getting.information.directly Overcoming.such.issues.is.challenging Therefore,.this.area.has.seen.some.attention,.but.is.relatively.unexplored.by.EDM.researchers

shop.referred.to.as.“Educational.data.mining”.occurring.in.2005 Since.then,.it.has.held.its.third.international.conference.in.2010,.had.one.book.published,.has.its.own.online.journal, and.is.now having this.book.published This.growth.is exciting for multiple.reasons First, education is a fundamentally important topic, rivaled only by medi-cine.and.health,.which.cuts.across.countries.and.cultures Being.able.to.better.answer.age-old.questions.in.education,.as.well.as.finding.ways.to.answer.questions.that.have.not.yet.been.asked,.is.an.activity.that.will.have.a.broad.impact.on.humanity Second,.doing.effective.educational.research.is.no.longer.about.having.a.large.team.of.graduate.assistants.to.score.and.code.data,.and.sufficient.offices.with.filing.cabinets.to.store.the.results There.are.public.repositories.of.educational.data.sets.for.others.to.try.their.hand.at.EDM,.and.anyone.with.a.computer.and.Internet.connection.can.join.the.community Thus,.a.much.larger.and.broader.population.can.participate.in.helping.improve.the.state.of.education

The.field.of.EDM.has.grown.substantially.in.the.past.five.years,.with.the.first.work-This.book.is.a.good.first.step.for.anyone.wishing.to.join.the.EDM.community,.or.for.active.researchers.wishing.to.keep.abreast.of.the.field The.chapters.are.written.by.key.EDM.researchers,.and.cover.many.of.the.field’s.essential.topics Thus,.the.reader.gets.a.broad.treatment.of.the.field.by.those.on.the.front.lines

MATLAB® is a registered trademark of The MathWorks, Inc For product information,.please.contact:

Dr Cristóbal Romero is an associate professor in the.Department.of.Computer.Science.at.the.University.of.Córdoba,.Spain His.research.interests.include.applying.artificial.intelli-gence.and.data-mining.techniques.in.education.and.e-.learning.systems He received his PhD in computer science from the.University of Granada, Spain, in 2003 The title of his PhD.thesis was “Applying data mining techniques for improving.adaptive hypermedia web-based courses.” He has published.several.papers.about.educational.data.mining.in.international.journals and conferences, and has served as a reviewer for.journals.and.as.a.program.committee.(PC).member.for.confer-ences He.is.a.member.of.the.International.Working.Group.on.Educational.Data.Mining.and.an.organizer.or.PC.member.of.conferences.and.workshops.about.EDM He.was.conference.chair.(with.Sebastián.Ventura).of.the.Second.International.Conference.on.Educational.Data.Mining.(EDM’09).

Dr Sebastián Ventura is an associate professor in the.Department of Computer Science at the University of.Córdoba, Spain He received his PhD in sciences from the.University.of.Córdoba.in.2003 His.research.interests.include.machine.learning,.data.mining,.and.their.applications,.and,.recently,.in.the.application.of.KDD.techniques.in.e-learning He.has.published.several.papers.about.educational.data.min-ing (EDM) in international journals and conferences He

has served as a reviewer for several journals such as User

Modelling and User Adapted Interaction , Information Sciences, and Soft Computing He.has.also.served.as.a PC.member.in.

several.research.EDM.forums,.including.as.conference.chair.(with Cristóbal Romero) of the Second International Conference on Educational Data.Mining.(EDM’09)

Dr Mykola Pechenizkiy is an assistant professor in the.Department of Computer Science, Eindhoven University of.Technology, the Netherlands He received his PhD in com-puter science and information systems from the University.of.Jyväskylä,.Finland,.in.2005 His.research.interests.include.knowledge discovery, data mining, and machine learning,.and.their.applications One.of.the.particular.areas.of.focus.is.on.applying.machine.learning.for.modeling,.changing.user.interests.and.characteristics.in.adaptive.hypermedia.applica-tions.including,.but.not.limited.to,.e-learning.and.e-health He.has.published.several.papers.in.these.areas,.and.has.been.involved.in.the.organization.of.conferences,.workshops,.and.special.tracks

Dr Ryan S J d Bakerogy and the learning sciences in the Department of Social.Science.and.Policy.Studies.at.Worcester.Polytechnic.Institute,.Massachusetts, with a collaborative appointment in com-puter.science He.graduated.from.Carnegie.Mellon.University,.Pittsburgh, Pennsylvania, in 2005, with a PhD in human–computer interaction He was a program chair (with Joseph.Beck) of the First International Conference on Educational.

.is.an.assistant.professor.of.psychol-Data Mining, and is an associate editor of the Journal of

Educational Data Mining and a founder of the International.Working.Group.on.Educational.Data.Mining His.research.is.at.the.intersection.of.educational.data.mining,.machine.learn-ing,.human–computer.interaction,.and.educational.psychology,.and.he.has.received.five.best.paper.awards.or.nominations.in.these.areas He.is.the.former.technical.director.of.the Pittsburgh Science of Learning DataShop, the world’s largest public repository for.data.on.the.interaction.between.students.and.educational.software

Wil van der Aalst

CharlotteCharlotte,.North.Carolina

Joseph E Beck

Computer.Science.DepartmentWorcester.Polytechnic.InstituteWorcester,.Massachusettsand

Machine.Learning.DepartmentCarnegie.Mellon.UniversityPittsburgh,.Pennsylvania

Javier Bravo

Escuela.Politécnica.SuperiorUniversidad.Autónoma.de.MadridMadrid,.Spain

Winslow Burleson

Department.of.Computer.Science.and.Engineering

Arizona.State.UniversityTempe,.Arizona

Toon Calders

Department.of.Mathematics.and

Computer.ScienceEindhoven.University.of.TechnologyEindhoven,.the.Netherlands

Cristina Carmona

Departamento.de.Lenguajes.y.Ciencias.de.la.Computación

Universidad.de.MálagaMálaga,.Spain

Kyle Cunningham

Human–Computer.Interaction.InstituteCarnegie.Mellon.University

Pittsburgh,.Pennsylvania

Sidney D’Mello

Institute.for.Intelligent.SystemsThe.University.of.MemphisMemphis,.Tennessee

Philippe Fournier-Viger

Department.of.Computer.ScienceUniversity.of.Quebec.in.MontrealMontreal,.Quebec,.Canada

Enrique García

Department.of.Computer.Science.and.Numerical.Analysis

University.of.CordobaCordoba,.Spain

Daniela Godoy

ISISTAN.Research.InstituteUniversidad.Nacional.del.Centro.de.la.Provincia.de.Buenos.Aires

Tandil,.Argentina

Evandro Gouvea

European.Media.Laboratory.GmbHHeidelberg,.Germany

and

Robotics.InstituteCarnegie.Mellon.UniversityPittsburgh,.Pennsylvania

Art Graesser

Institute.for.Intelligent.SystemsThe.University.of.MemphisMemphis,.Tennessee

Kenneth R Koedinger

Human–Computer.Interaction.InstituteCarnegie.Mellon.University

Pittsburgh,.Pennsylvania

Irena Koprinska

School.of.Information.TechnologiesUniversity.of.Sydney

Sydney,.New.South.Wales,.Australia

Brett Leber

Human–Computer.Interaction.InstituteCarnegie.Mellon.University

Pittsburgh,.Pennsylvania

Tara M Madhyastha

Department.of.PsychologyUniversity.of.WashingtonSeattle,.Washington

David Masip

Department.of.Computer.Science,

Multimedia.and.TelecommunicationsUniversitat.Oberta.de.Catalunya

Barcelona,.Spain

Manolis Mavrikis

London.Knowledge.LabThe.University.of.LondonLondon,.United.Kingdom

Riccardo Mazza

Faculty.of.Communication.SciencesUniversity.of.Lugano

Lugano,.Switzerlandand

Department.of.Innovative.TechnologiesUniversity.of.Applied.Sciences.of.Southern.Switzerland

Manno,.Switzerland

John C Nesbit

Faculty.of.EducationSimon.Fraser.UniversityBurnaby,.British.Columbia,.Canada

Engelbert Mephu Nguifo

Department.of.Computer.SciencesUniversité.Blaise-Pascal.Clermont.2Clermont-Ferrand,.France

Roger Nkambou

Department.of.Computer.ScienceUniversity.of.Quebec.in.MontrealMontreal,.Quebec,.Canada

Alvaro Ortigosa

Escuela.Politécnica.SuperiorUniversidad.Autónoma.de.MadridMadrid,.Spain

Zachary A Pardos

Department.of.Computer.ScienceWorcester.Polytechnic.InstituteWorcester,.Massachusetts

Mykola Pechenizkiy

Department.of.Mathematics.and

Computer.ScienceEindhoven.University.of.TechnologyEindhoven,.the.Netherlands

Kaska Porayska-Pomsta

London.Knowledge.LabThe.University.of.LondonLondon,.United.Kingdom

Manuel E Prieto

Escuela.Superior.de.InformáticaUniversidad.de.Castilla-La.ManchaCiudad.Real,.Spain

Alfredo Vellido

Departament.de.Llenguatges.i.Sistemes.Informàtics

Universitat.Politècnica.de.CatalunyaBarcelona,.Spain

Sebastián Ventura

Department.of.Computer.Science.and.Numerical.Analysis

University.of.CordobaCordoba,.Spain

César Vialardi

Facultad.de.Ingeniería.de.SistemasUniversidad.de.Lima

Lima,.Peru

Mikko Vinni

School.of.ComputingUniversity.of.Eastern.FinlandJoensuu,.Finland

Stephan Weibelzahl

School.of.ComputingNational.College.of.IrelandDublin,.Ireland

Philip H Winne

Faculty.of.EducationSimon.Fraser.UniversityBurnaby,.British.Columbia,.Canada

Beverly P Woolf

Department.of.Computer.ScienceUniversity.of.Massachusetts.AmherstAmherst,.Massachusetts

Mingming Zhou

Faculty.of.EducationSimon.Fraser.UniversityBurnaby,.British.Columbia,.Canada

its.third.iteration),.a.journal.(the.Journal of Educational Data Mining),.and.a.number.of.highly.

cited.papers.(see.[2].for.a.review.of.some.of.the.most.highly.cited.EDM.papers)

These contributions in education build off of data mining’s past impacts in other.domains.such.as.commerce.and.biology.[11] In.some.ways,.the.advent.of.EDM.can.be.con-sidered.as.education.“catching.up”.to.other.areas,.where.improving.methods.for.exploiting.data have promoted transformative impacts in practice [4,7,12] Although the discovery.methods.used.across.domains.are.similar.(e.g [3]),.there.are.some.important.differences.between.them For.instance,.in.comparing.the.use.of.data.mining.within.e-commerce.and.EDM,.there.are.the.following.differences:

CONTENTS

1.1 Background 11.2 Educational.Applications 31.3 Objectives,.Content,.and.How.to.Read.This.Book 4References 5

•

Domain The.goal.of.data.mining.in.e-commerce.is.to.influence.clients.in.purchas-ing.while.the.educational.systems.purpose.is.to.guide.students.in.learning.[10]

• Data In e-commerce, typically data used is limited to web server access logs,.

whereas.in.EDM.there.is.much.more.information.available.about.the.student.[9],.allowing.for.richer.user.(student).modeling This.data.come.possibly.from.differ-ent.sources,.including.field.observations,.motivational.questionnaires,.measure-ments.collected.from.controlled.experiments,.and.so.on Depending.on.the.type.of.the.educational.environment.(traditional.classroom.education,.computer-based.or.web-based.education).and.an.information.system.that.supports.it.(a.learning.management,.an.intelligent.tutoring.or.adaptive.hypermedia.system).also.differ-ent.kinds.of.data.is.being.collected.including.but.not.limited.to.student.profiles,.(inter)activity.data,.interaction.(with.the.system,.with.educators.and.with.peers),.rich.information.about.learning.objects.and.tasks,.and.so.on Gathering.and.inte-grating.this.data.together,.performing.its.exploratory.analysis,.visualization,.and.preparation.for.mining.are.nontrivial.tasks.on.their.own

• Objective The.objective.of.data.mining.in.e-commerce.is.increasing.profit Profit.

is.a.tangible.goal.that.can.be.measured.in.terms.of.amounts.of.money,.and.which.leads.to.clear.secondary.measures.such.as.the.number.of.customers.and.customer.loyalty As.the.objective.of.data.mining.in.education.is.largely.to.improve.learning.[10],.measurements.are.more.difficult.to.obtain,.and.must.be.estimated.through.proxies.such.as.improved.performance

• Techniques The.majority.of.traditional.data.mining.techniques.including.but.not.

limited.to.classification,.clustering,.and.association.analysis.techniques.have.been.already.applied.successfully.in.the.educational.domain And.the most popular.approaches.are.covered.by.the.introductory.chapters.of.the.book Nevertheless,.educational.systems.have.special.characteristics.that.require.a.different.treatment.of.the.mining.problem Data.hierarchy.and.nonindependence.becomes.particu-larly important to account for, as.individual students contribute large amounts.of.data.while.progressing.through.a.learning.trajectory,.and.those.students.are.impacted.by.fellow.classmates.and.teacher.and.school-level.effects As.a.conse-quence, some specific data mining techniques are needed to address learning.[8] and other data about learners Some traditional techniques can be adapted,.some.cannot This.trend.has.led.to.psychometric.methods.designed.to.address.these.issues.of.hierarchy.and.nonindependence.being.integrated.into.EDM,.as.can.be.seen.in.several.chapters.in.this.volume However,.EDM.is.still.an.emerging.research area, and we can foresee that its further development will result in a.better.understanding.of.challenges.peculiar.to.this.field.and.will.help.researchers.involved.in.EDM.to.see.what.techniques.can.be.adopted.and.what.new.tailored.techniques.have.to.be.developed

The application of data mining techniques to educational systems in order to improve.learning.can.be.viewed.as.a.formative.evaluation.technique Formative.evaluation.[1].is.the.evaluation.of.an.educational.program.while.it.is.still.in.development,.and.with.the.purpose.of.continually.improving.the.program Examining.how.students.use.the.system.is.one.way.to.evaluate.instructional.design.in.a.formative.manner.and.may.help.educa-tional.designers.to.improve.the.instructional.materials.[5] Data.mining.techniques.can.be.used.to.gather.information.that.can.be.used.to.assist.educational.designers.to.establish.a

pedagogical.basis.for.decisions.when.designing.or.modifying.an.environment’s.pedagogi-The.application.of.data.mining.to.the.design.of.educational.systems.is.an.iterative.cycle.of.hypothesis.formation,.testing,.and.refinement.(see.Figure.1.1)

Mined knowledge should enter the design loop towards guiding, facilitating, and.enhancing.learning.as.a.whole In.this.process,.the.goal.is.not.just.to.turn.data.into.knowl-edge,.but.also.to.filter.mined.knowledge.for.decision.making

As we can see in Figure 1.1, educators and educational designers (whether in school.districts,.curriculum.companies,.or.universities).design,.plan,.build,.and.maintain.educa-tional.systems Students.use.those.educational.systems.to.learn Building.off.of.the.avail-able.information.about.courses,.students,.usage,.and.interaction,.data.mining.techniques.can.be.applied.in.order.to.discover.useful.knowledge.that.helps.to.improve.educational.designs The.discovered.knowledge.can.be.used.not.only.by.educational.designers.and.teachers,.but.also.by.end.users—students Hence,.the.application.of.data.mining.in.educa-tional.systems.can.be.oriented.to.supporting.the.specific.needs.of.each.of.these.categories.of.stakeholders

1.2 Educational Applications

In.the.last.several.years,.EDM.has.been.applied.to.address.a.wide.number.of.goals In.this.book.we.can.distinguish.between.the.following.general.applications.or.tasks:

• Communicating to stakeholders The.objective.is.to.help.to.course.administrators.and.

educators.in.analyzing.students’.activities.and.usage.information.in.courses The.most.frequently.used.techniques.for.this.type.of.goal.are.exploratory.data.analy-sis.through.statistical.analysis.and.visualizations.or.reports,.and.process.mining

Educational systems (traditional classrooms, e-learning systems, LMSs, web-based adaptive systems, intelligent tutoring systems, questionnaires and quizzes)

Provide, store:

Course information, contents, academic data, grades, student usage and interaction data Data mining techniques

(statistics, visualization, clustering, classification, association rule mining, sequence mining, text

mining)

Model learners and learning, communicate findings, make recommendations

FIGURE 1.1

Applying.data.mining.to.the.design.of.educational.systems.

• Maintaining and improving courses The.objective.is.to.help.to.course.administrators.

and.educators.in.determining.how.to.improve.courses.(contents,.activities,.links,.etc.),.using.information.(in.particular).about.student.usage.and.learning The.most.frequently.used.techniques.for.this.type.of.goal.are.association,.clustering,.and.classification Chapters.7,.17,.26,.and.34.discuss.methods.and.case.studies.for.this.category.of.application

• Generating

recommendation The.objective.is.to.recommend.to.students.which.con-tent.(or.tasks.or.links).is.most.appropriate.for.them.at.the.current.time The.most.frequently.used.techniques.for.this.type.of.goal.are.association,.sequencing,.clas-sification,.and.clustering Chapters.6,.8,.12,.18,.19,.and.32.discuss.methods.and.case.studies.for.this.category.of.application

• Predicting student grades and learning outcomes The.objective.is.to.predict.a.student’s.

final.grades.or.other.types.of.learning.outcomes.(such.as.retention.in.a.degree.program.or.future.ability.to.learn),.based.on.data.from.course.activities The.most.frequently.used.techniques.for.this.type.of.goal.are.classification,.clustering,.and.association Chapters.5.and.13.discuss.methods.and.case.studies.for.this.category.of.application

• Student

modeling User.modeling.in.the.educational.domain.has.a.number.of.appli-cations,.including.for.example.the.detection.(often.in.real.time).of.student.states.and.characteristics.such.as.satisfaction,.motivation,.learning.progress,.or.certain.types of problems that negatively impact their learning outcomes (making too.many.errors,.misusing.or.underusing.help,.gaming.the.system,.inefficiently.explor-ing.learning.resources,.etc.),.affect,.learning.styles,.and.preferences The.common.objective.here.is.to.create.a.student.model.from.usage.information The.frequently.used.techniques.for.this.type.of.goal.are.not.only.clustering,.classification,.and.association analysis, but also statistical analyses, Bayes networks (including.Bayesian.Knowledge-Tracing),.psychometric.models,.and.reinforcement.learning Chapters.6,.12,.14.through.16,.20,.21,.23,.25,.27,.31,.33,.and.35.discuss.methods.and.case.studies.for.this.category.of.application

• Domain structure analysis The.objective.is.to.determine.domain.structure,.using.

the.ability.to.predict.student.performance.as.a.measure.of.the.quality.of.a.domain.structure.model Performance.on.tests.or.within.a.learning.environment.is.uti-lized.for.this.goal The.most.frequently.used.techniques.for.this.type.of.goal.are.association.rules,.clustering.methods,.and.space-searching.algorithms Chapters.10,.11,.29,.and.30.discuss.methods.and.case.studies.for.this.category.of.application

1.3 Objectives, Content, and How to Read This Book

Our.objective,.in.compiling.this.book,.is.to.provide.as.complete.as.possible.a.picture.of.the current state of the art in the application of data mining techniques in education Recent.developments.in.technology.enhanced.learning.have.resulted.in.a.widespread.use

of e-learning environments and educational software, within many regular university

This.expansion.of.data.has.led.to.increasing.interest.among.education.researchers.in.a.variety.of.disciplines,.and.among.practitioners.and.educational.administrators,.in.tools.and.techniques.for.analysis.of.the.accumulated.data.to.improve.understanding.of.learners.and.learning.process,.to.drive.the.development.of.more.effective.educational.software.and.better.educational.decision-making This.interest.has.become.a.driving.force.for.EDM We.believe.that.this.book.can.support.researchers.and.practitioners.in.integrating.EDM.into.their.research.and.practice,.and.bringing.the.educational.and.data.mining.communities.together,.so.that.education.experts.understand.what.types.of.questions.EDM.can.address,.and data miners understand what types of questions are of importance to educational.design.and.educational.decision-making

This.volume,.the.Handbook of Educational Data Mining,.consists.of.two.parts In.the.first.

part,.we.offer.nine.surveys.and.tutorials.about.the.principal.data.mining.techniques.that.have.been.applied.in.education In.the.second.part,.we.give.a.set.of.25.case.studies,.offering.readers.a.rich.overview.of.the.problems.that.EDM.has.produced.leverage.for

The.book.is.structured.so.that.it.can.be.read.in.its.entirety,.first.introducing.concepts.and.methods,.and.then.showing.their.applications However,.readers.can.also.focus.on.areas.of.specific.interest,.as.have.been.outlined.in.the.categorization.of.the.educational.applications We.welcome.readers.to.the.field.of.EDM.and.hope.that.it.is.of.value.to.their.research.or.practical.goals If.you.enjoy.this.book,.we.hope.that.you.will.join.us.at.a.future.iteration.of.the.Educational.Data.Mining.conference;.see.www.educationaldatamining.org.for.the.latest.information,.and.to.subscribe.to.our.community.mailing.list,.edm-announce

future.visions Journal of Educational Data Mining,.1(1),.3–17.

3 Hanna,.M (2004) Data.mining.in.the.e-learning.domain Computers and Education Journal,.42(3),.

7 Lewis,.M (2004) Moneyball: The Art of Winning an Unfair Game New.York:.Norton.

8 Li, J and Zạane, O (2004) Combining usage, content, and structure data to improve web.

site.recommendation In.International Conference on Ecommerce and Web Technologies,.Zaragoza,.

Spain,.pp 305–315.

9 Pahl,.C and.Donnellan,.C (2003) Data.mining.technology.for.the.evaluation.of.web-based.

teaching.and.learning.systems In.Proceedings of the Congress e-Learning,.Montreal,.Canada 10 Romero,.C and.Ventura,.S (2007) Educational.data.mining:.A.survey.from.1995.to.2005 Expert Systems with Applications,.33(1),.135–146.

Basic Techniques, Surveys

and Tutorials

Riccardo Mazza

2.1 Introduction

This.chapter.presents.an.introduction.to.information.visualization,.a.new.discipline.with.origins.in.the.late.1980s.that.is.part.of.the.field.of.human–computer.interaction We.will.illustrate.the.purposes.of.this.discipline,.its.basic.concepts,.and.some.design.principles.that.can.be.applied.to.graphically.render.students’.data.from.educational.systems The.chapter starts with.a description.of information visualization.followed by a.discussion.on.some.design.principles,.which.are.defined.by.outstanding.scholars.in.the.field Finally,.some.systems.in.which.visualizations.have.been.used.in.learning.environments.to.repre-sent.user.models,.discussions,.and.tracking.data.are.described

CONTENTS

2.1 Introduction 92.2 What.Is.Information.Visualization? 102.2.1 Visual.Representations 102.2.2 Interaction 112.2.3 Abstract.Data 112.2.4 Cognitive.Amplification 122.3 Design.Principles 132.3.1 Spatial.Clarity 142.3.2 Graphical.Excellence 142.4 Visualizations.in.Educational.Software 162.4.1 Visualizations.of.User.Models 162.4.1.1 UM/QV 162.4.1.2 ViSMod 172.4.1.3 E-KERMIT 182.4.2 Visualizations.of.Online.Communications 192.4.2.1 Simuligne 192.4.2.2 PeopleGarden 202.4.3 Visualizations.of.Student-Tracking.Data 202.5 Conclusions 24References 25

2.2 What Is Information Visualization?

Visualization,.which.may be.defined.as.“the display.of.data.with the.aim.of ing.comprehension.rather.than.photographic.realism*”,.has.greatly.increased.over.the.last.years.thanks.to.the.availability.of.more.and.more.powerful.computers.at.low.cost The.dis-cipline.of.information.visualization.(IV).[2,16].originated.in.the.late.1980s.for.the.purpose.of.exploring.the.use.of.computers.to.generate.interactive,.visual.representation.to.explain.and.understand.specific.features.of.data The.basic.principle.of.IV.is.to.present.data.in.a.visual.form.and.use.human.perceptual.abilities.for.their.interpretation

maximiz-As.in.many.other.fields,.several.people.have.tried.to.give.a.rigorous,.scientific.definition.of.the.discipline.of.IV The.definition.that.received.most.consensus.from.the.community.of.the.researchers.seems.to.be.the.one.given.by.Card.et.al in.their.famous.collection.of.papers

on.IV:.the

readings.[2] According.to.them,.IV.is.“the.use.of.computer-supported,.interac-tive,.visual.representations.of.abstract.data.to.amplify.cognition.”.By.this.definition,.four.terms.are.the.key.to.understand.this.domain:.visual.representation,.interaction,.abstract.data,.and.cognitive.amplification We.will.try.to.analyze.each.of.them.to.clearly.describe.the.field.and.their.applications

2.2.1 Visual Representations

ena,.data,.and.events.using.graphics Some.aspects,.such.as.when.people.need.to.find.a.route.in.a.city,.the.stock.market.trends.over.a.certain.period,.and.the.weather.forecast,.may.be.understood.better.using.graphics.rather.than.text Graphical.representation.of.data,.compared.to.the.textual.or.tabular.ones.(in.case.of.numbers),.takes.advantage.of.the.human.visual.perception Perception.is.very.powerful.as.it.conveys.large.amount.of.information.to.our.mind,.and.allowing.us.to.recognize.essential.features.and.to.make.important.inferences This.is.possible.thanks.to.the.fact.that.there.is.a.series.of.identifica-tion.and.recognition.operations.that.our.brain.performs.in.an.“automatic”.way.without.the.need.to.focus.our.attention.or.even.be.conscious.of.them Perceptual.tasks.that.can.be.performed.in.a.very.short.time.lapse.(typically.between.200.and.250.ms.or.less).are.called.pre-attentive,.since.they.occur.without.the.intervention.of.consciousness.[20]

bars whose length is proportional to the number on the left

Suppose we have to find the maximum and the minimum

FIGURE 2.1

Comparing perception of lines.with.numbers.

Let.us.try.to.do.the.same.operation,.this.time.using.the.bars.on.the.left The.length.of.the.bars.lets.us.to.identify.almost.immediately.the.longest.and.the.shortest.thanks.to.the.pre-attentive.property.of.length,.the.length.of.the.bars.allows.us.to.almost.immediately.identify.the.longest.and.the.shortest

alisation”.in.the.British.version.of.the.term) It.has.been.noted.by.Spence.[16].that.there.is.a.diversity.of.uses.of.the.term.“visualization.”.For.instance,.in.a.dictionary.the.following.definitions.can.be.found:

Graphical.representations.are.often.associated.with.the.term.“visualization”.(or.“visu-Visualize:.form.a.mental.image.of…*

Visualization:.The.display.of.data.with.the.aim.of.maximizing.comprehension.rather than.photographic.realism †

Visualization:.the.act.or.process.of.interpreting.in.visual.terms.or.of.putting.into.visible form ‡

These.definitions.reveal.that.visualization.is.an.activity.in.which.humans.are.engaged,.as.an.internal.construct.of.the.mind.[16,20] It.is.something.that.cannot.be.printed.on.a.paper

or displayed on a computer screen With these considerations, we can summarize that.visualization.is.a.cognitive.activity,.facilitated.by.graphical.external.representations.from.which.people.construct.internal.mental.representation.of.the.world.[16,20]

Computers.may.facilitate.the.visualization.process.with.some.visualization.tools This.is.especially.true.in.recent.years.with.the.availability.of.powerful.computers.at.low.cost However,.the.above.definition.is.independent.from.computers:.although.computers.can.facilitate.visualization,.it.still.remains.an.activity.that.happens.in.the.mind

2.2.2 Interaction

Recently.there.has.been.great.progress.in.high-performance,.affordable.computer.graphics The.common.personal.computer.has.reached.a.graphic.power.that.just.10.years.ago.was.possible.only.with.very.expensive.graphic.workstations.specifically.built.for.the.graphic.process At.the.same.time,.there.has.been.a.rapid.expansion.in.information.that.people.have.to.process.for.their.daily.activities This.need.led.scientists.to.explore.new.ways.to.represent.huge.amounts.of.data.with.computers,.taking.advantage.of.the.possibility.of.users.interact-ing.with.the.algorithms.that.create.the.graphical.representation Interactivity.derives.from.the.people’s.ability.to.also.identify.interesting.facts.when.the.visual.display.changes.and.allows.them.to.manipulate.the.visualization.or.the.underlying.data.to.explore.such.changes

2.2.3 Abstract Data

IV.definitions.introduce.the.term.“abstract.data,”.for.which.some.clarification.is.needed The.data.itself.can.have.a.wide.variety.of.forms,.but.we.can.distinguish.between.data.that.have.a.physical.correspondence.and.is.closely.related.to.mathematical.structures.and.models.(e.g.,.the.airflow.around.the.wing.of.an.airplane,.or.the.density.of.the.ozone.layer.surrounding

* The Concise Oxford Dictionary Ed Judy.Pearsall Oxford.University.Press,.2001 Oxford Reference Online Oxford.

University.Press.

†.A Dictionary of Computing Oxford.University.Press,.1996 Oxford Reference Online Oxford.University.Press.

in Figure 2.2), while IV is dealing with unstructured data.

sets as a distinct flavor [4] In Table 2.1 is reported a table

2.2.4 Cognitive Amplification

plication.(a.typical.mental.activity),.e.g.,.27.×.42.in.our.head,.without.having.a.pencil.and.paper This.calculation.made.with.our.mind.will.take.usually.at.least.five.times.longer.than.when.using.a.pencil.and.paper.[2] The.difficulty.in.doing.this.operation.in.the.mind.is.holding.the.partial.results.of.the.multiplication.in.the.memory.until.they.can.be.used:

FIGURE 2.2

Example of scientific tion: The ozone hole the South Pole on September 22, 2004 (Image.from.the.NASA.Goddard Space.Center.archives.and.repro- duced.with.permission.)

for.patterns,.recognize.relationship.between.data,.and.perform.some.inferences.more.easily Card.et.al [2].propose.six.major.ways.in.which.visualizations.can.amplify.cognition.by 1 Increasing.the.memory.and.processing.resources.available.to.users

is a well-known old adage that everybody knows But why (and in which situations).graphical.representations.are.effective?

SH

ZH

ZG BS

AR AI

Abstimmung vom 24 September 2006 Votation du 24 Septembre 2006

Änderung des Asylgesetzes

Modification de la loi sur l΄asile

Schweizerische Eidgenossenschaft Eidgenössisches Departement des Innern EDI

Département fédéral de l΄intérieur DFI Bundesamt für Statistik BFS Confédération suisse

Confederazione svizzera

Confederaziun svizra Office fédéral de la statistique OFS

Quelle: Abstimmungsstatistik, BFS Source: Statistique des votations, OFS

© BFS, ThemaKart, Neuenburg 2006/K17.A525.R_bz

© OFS, ThemaKart, Neuchâtel 2006/K17.A525.R_bz

FIGURE 2.3

Graphical.representation.of.results.of.federal.referendum.in.Switzerland.on.September.24,.2006 (Image.from the Swiss Federal Statistical Office, http://www.bfs.admin.ch © Bundesamt für Statistik, ThemaKart 2009, reproduced.with.permission.)

2.3.1 Spatial Clarity

Graphical.representations.may.facilitate.the.way.we.present.and.understand.large complex.datasets As.Larkin.and.Simon.[7].argued.in.their.seminal.paper.“Why.a.diagram.is.(some-times).worth.ten.thousand.words,”.the.effectiveness.of.graphical.representations.is.due

to their spatial clarity Well-constructed graphical representations of data allow people.to.quickly.to.gain.insights.that.might.lead.to.significant.discoveries.as.a.result.of.spatial.clarity

Larkin.and.Simon.compared.the.computational.efficiency.of.diagrams.and.sentences.in.solving.physics.problems,.and.concluded.that.diagrams.helped.in.three.basic.ways:

Locality.is.enabled.by.grouping.together.information.that.is.used.together This.avoids.large.amounts.of.search.and.allows.different.information.closely.located.to.be.processed.simulta-neously For.example,.Figure.2.4.represents.the.map.of.the.Madrid.metro.transport.system In.this.map.the.locality.principle.is.applied.by.placing.metro.lines.and.zones.in.the.same.map The.traveler.can.find.in.the.same.place.information.about.lines,.connections,.and.stations

Minimizing labeling.is.enament,.avoiding.the.need.to.match.symbolic.labels.and.leading.to.reducing.the.working.memory load For example, the Madrid transport map (Figure 2.4) uses visual entities.such.as.lines.depicted.with.different.colors.to.denote.different.metro.lines Connections.are.clearly.indicated.by.a.white.circle.that.connects.the.corresponding.lines There.is.no.need.to.use.textual.representations.because.the.connections.are.explicitly.represented.in.the.graphics

bled.by.using.location.to.group.information.about.a.single.ele-Perceptual enhancement.is.enabled.by.supporting.a.large.number.of.perceptual.inferences.that.are.easy.for.humans.to.perform For.example,.in.Figure.2.4,.a.traveler.who.has.to.travel

from.Nuevos Ministerios.to.Opera.can.see.that.there.are.different.combination.of.lines.and.

connection.that.he.can.take,.and.probably.can.decide.which.is.the.fastest.way.to.reach.the.destination

2.3.2 Graphical Excellence

Sometimes graphical representations of data have been used to distort the underlying.data Tufte.[18].and.Bertin.[1].list.a.number.of.examples.of.graphics.that.distort.the.under-lying data or communicate incorrect ideas Tufte indicates some principles that should.be.followed.to.build.effective.well-designed.graphics In.particular,.a.graphical.display.should.[18]

Metro de Madrid

© 2007 Designed and drawn by Matthew McLauchlin, http://www.metrodemontreal.com/

This version released under Creative Commons Share-Alike Attribution Licence (CC-SA-BY 2.5)

Europa 12

Arganzuela-Méndez Álvardo Puente de Vallecas

Rivas Urbanizaciones Rivas Vaciamadrid

La Poveda

Nueva Numancia Portazgo Buenos Aires Alto del Arenal Miguel Hernández Sierra de Guadalupe Villa de Vallecas Congosto

La Gavia Las Suertes Valdecarros

1

Almendrales Ciudad de los Ángeles Villaverde Bajo-Cruce Villaverde Alto

3San Cristóbal

San Fermin-Orcasur Hospital 12 de Octubre Usera

Urgel Pirámides Laguna

Lucero Pta de

Toledo

La Latina

Atocha Renfe

Tirso de Molina Antón Martin Atocha

Banco de España

Conde de Casal

Ibiza

O΄Donnell

Ascao Quintana

El Carmen

Barrio de la Concepción

Pque de las Avenidas ProsperidadAlfonso XIIIAvenida de la Paz

Canillas Esperanza

Campo de las Naciones

Barajas Aeropuerto T1-T2-T3

Arturo Soria

El Capricho Canillejas Torre Arias Suanzes Ciudad Lineal Cartagena

Lista Velázquez Colón

Rubén Dario

Iglesia Quevedo

Islas

Filipinas

Francos RodriguezValdezarza Alvarado

Estrecho Santiago Bernabéu Tetuán

Valdeacederas

Plaza de Castilla

Barrio del Pilar Ventilla

Las Tablas

La Granja Ronda de la Comunicación

Cuzco

Duque de Pastrana Pío XII Bambú

República Argentina

Concha Espina

Hortaleza Pinar del Rey

Manoteras

Fuencarral Begoña

Marqués

de la Valdavia Baunatal

Reyes Católicos

Manuel de Falla

La Moraleja

Parque de Santa María San lorenzo

Cruz del Rayo

Antonio MachadoPeñagrande

Avenida de la IllustraciónLacoma

Arroyo del Fresno

Metropolitano

Cano Rios Rosas

Serrano

Retiro Chueca Sevilla

Garcia Noblejas Simancas San Blas Las Musas Barrio del Puerto Coslada Central

La Rambla San Fernando Jarama Lavapiés

Menéndez Pelayo Palos de la Frontera

Plaza Eliptica

La Peseta

11

Legazpi 11

Oporto Acacias

Ópera 5

Alonso Martínez Príncipe de Vergara

Tribunal Gran Via Callao

Bilbao Noviciado Pza de España

Príncipe Pío

Núñez

de Balboa

La Elipa Ventas

Diego de León Avda de América

Hospital del Norte

10

8 Pinar de chamartin

Tres Olivos

9

Herrera Oria Pitis

6

Alameda

de Osuna 5

2

Pueblo Nuevo

6 Pacífico

Puerta de Arganda

Henares7

Estrella Vinateros Artilleros Pavones Valdebernardo Vicálvaro San Cipriano

Arganda del Rey

FIGURE 2.4

Map.of.the.Madrid.metro.system (Images.licensed.under.Creative.Commons.Share-Alike.)

A.key.question.in.IV.is.how.we.convert.abstract.data.into.a.graphical.representation,.preserving.the.underlying.meaning.and,.at.the.same.time,.providing.new.insight There.is.no.“magic.formula”.that.helps.the.researchers.to.build.systematically.a.graphical.repre-sentation.starting.from.a.raw.set.of.data It.depends.on.the.nature.of.the.data,.the.type.of.information.to.be.represented.and.its.use,.but.more.consistently,.it.depends.on.the.creativ-ity.of.the.designer.of.the.graphical.representation Some.interesting.ideas,.even.if.innova-tive,.have.often.failed.in.practice

Graphics.facilitate.IV,.but.a.number.of.issues.must.be.considered.[16,18]:

1 Data is nearly always multidimensional, while graphics represented on a puter.screen.or.on.a.paper.are.presented.in.a.2D.surface

com- 2com-.com-.Sometimescom-.wecom-.needcom-.tocom-.representcom-.acom-.hugecom-.dataset,com-.whilecom-.thecom-.numbercom-.ofcom-.datacom-.view-able.on.a.computer.screen.or.on.a.paper.is.limited

2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view- 32 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.Data2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.may2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.vary2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.during2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.the2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.time,2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.while2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.graphics2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.are2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-.static2 Sometimes.we.need.to.represent.a.huge.dataset,.while.the.number.of.data.view-

4 Humans.have.remarkable.abilities.to.select,.manipulate,.and.rearrange.data,.so.the.graphical.representations.should.provide.users.with.these.features

2.4 Visualizations in Educational Software

In.this.section.we.will.explore.some.graphical.representations.that.have.been.adopted.in.educational.contexts We.will.concentrate.our.analysis.in.software.applications.that.aims.to.provide.learning.to.students.and.gives.the.instructors.some.feedback.on.actions.and.improvements.undertaken.by.students.with.the.subject We.will.consider.three.types.of.applications:.visualization.of.user.models,.visualization.of.online.communications,.and.visualization.of.students’.tracking.data

2.4.1 Visualizations of User Models

edge.in.various.areas,.and.their.goals.and.preferences Student.models.are.a.key.compo-nent.of.intelligent.educational.systems.used.to.represent.the.student’s.understanding.of.material.taught Methods.for.user.modeling.are.often.exploited.in.educational.systems These.models.are.enabling.the.increasing.personalization.of.software,.particularly.on.the.Internet,.where.the.user.model.is.the.set.of.information.and.beliefs.that.is.used.to.person-alize.the.Web.site.[19]

A.user.model.is.a.representation.of.a.set.of.beliefs.about.the.user,.particularly.their.knowl-2.4.1.1  UM/QV

QV.[6].is.an.overview.interface.for.UM.[5],.a.toolkit.for.cooperative.user.modeling A.model.is.structured.as.a.hierarchy.of.elements.of.the.domain QV.uses.a.hierarchical.representa-tion.of.concepts.to.present.the.user.model For.instance,.Figure.2.5.gives.a.graphical.rep-resentation.of.a.model.showing.concepts.of.the.SAM.text.editor It.gives.a.quick.overview.whether.the.user.appears.to.know.each.element.of.the.domain QV.exploits.different.types

of.geometric.forms.and.color.to.represent.known/unknown.concepts A.square.indicates.a.knowledge.component,.diamond.a.belief,.a.circle.indicates.a.nonleaf.node,.and.crosses.indi-cate.other.component.types The.filling.of.the.shape.is.used.to.indicate.the.component.value For.instance,.in.the.example,.the.white.squares.show.that.the.user.knows.that.element,.while.the.dark.squares.indicate.lack.of.knowledge Nested.shapes,.such.as.default _ size _ k.

or undo _ k,.indicate.that.the.system.has.not been.able to.determine.whether the user.knows.it.or.not.(e.g.,.if.there.is.inconsistency.in.the.information.about.the.user) The.view.of.the.graph.is.manipulable,.in.particular,.clicking.on.a.nonleaf.node.causes.the.subtree.to.be.displayed,.useful.in.case.of.models.having.a.large.number.of.components.to.be.displayed

2.4.1.2  ViSMod

ViSMod.[22].is.an.interactive.visualization.tool.for.the.representation.of.Bayesian.learner.models In ViSMod, learners and instructors can inspect the learner model using a.graphical.representation.of.the.Bayesian.network ViSMod.uses.concept.maps.to.render.a

Quit overview

Select node

to fold

or unfold

Label all nodes

Minimal quit_k quit_b default_size_k non_cmd_b undo_k load_new_k

set_fname_k more_useful

Sam Editors

Root

write_k Mouse

Other command_window exch_k search_k xerox_k Powerful

mostly_useless emacs

vi c_c pascal_c Programming

languages

lisp_c fortran_c typing_ok_c user_info

Useful Basics

Label leaves only

No labels

very_useful mouse

command_window gotoline_k

FIGURE 2.5

The.QV.tool.showing.a.user.model (Image.courtesy.of.Judy.Kay.)