Download Full Handbook of educational data mining 1st Edition Cristobal Romero PDF All Chapters

Download the full version of the ebook at
https://ebookultra.com
Handbook of educational data mining 1st

Edition Cristobal Romero
https://ebookultra.com/download/handbook-of-
educational-data-mining-1st-edition-cristobal-
romero/
Explore and download more ebook at https://ebookultra.com

Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.
The Handbook of Data Mining 1st Edition Nong Ye
https://ebookultra.com/download/the-handbook-of-data-mining-1st-
edition-nong-ye/
ebookultra.com
Handbook of Statistical Analysis and Data Mining

Applications 1st Edition Robert Nisbet
https://ebookultra.com/download/handbook-of-statistical-analysis-and-
data-mining-applications-1st-edition-robert-nisbet/
ebookultra.com
Data Mining and Data Warehousing 1st Edition S.K. Mourya
https://ebookultra.com/download/data-mining-and-data-warehousing-1st-
edition-s-k-mourya/
ebookultra.com
Exploratory Data Mining and Data Cleaning 1st Edition

Tamraparni Dasu
https://ebookultra.com/download/exploratory-data-mining-and-data-
cleaning-1st-edition-tamraparni-dasu/
ebookultra.com
Biological Data Mining Chapman Hall Crc Data Mining and
Knowledge Discovery Series 1st Edition Jake Y. Chen
https://ebookultra.com/download/biological-data-mining-chapman-hall-
crc-data-mining-and-knowledge-discovery-series-1st-edition-jake-y-
chen/
ebookultra.com
Music Data Mining 1st Edition Tao Li
https://ebookultra.com/download/music-data-mining-1st-edition-tao-li/
ebookultra.com
Data Mining for Genomics and Proteomics Analysis of Gene

and Protein Expression Data Wiley Series on Methods and
Applications in Data Mining 1st Edition D. M. Dziuda
https://ebookultra.com/download/data-mining-for-genomics-and-
proteomics-analysis-of-gene-and-protein-expression-data-wiley-series-
on-methods-and-applications-in-data-mining-1st-edition-d-m-dziuda/
ebookultra.com
Data Mining for Bioinformatics Applications 1st Edition He

Zengyou
https://ebookultra.com/download/data-mining-for-bioinformatics-
applications-1st-edition-he-zengyou/
ebookultra.com
Practical Data Mining 1st Edition Monte F. Hancock Jr
https://ebookultra.com/download/practical-data-mining-1st-edition-
monte-f-hancock-jr/
ebookultra.com
Handbook of educational data mining 1st Edition
Cristobal Romero Digital Instant Download
Author(s): Cristobal Romero, Sebastian Ventura, Mykola Pechenizkiy, Ryan
S.J.d. Baker
ISBN(s): 9781439804582, 1439804583
Edition: 1
File Details: PDF, 13.87 MB
Year: 2010
Language: english
Handbook of
Educational Data Mining
Chapman & Hall/CRC
Data Mining and Knowledge Discovery Series
SERIES EDITOR
Vipin Kumar
University 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.
PUBLISHED TITLES
UNDERSTANDING COMPLEX DATASETS: TEXT MINING: CLASSIFICATION, CLUSTERING,
DATA MINING WITH MATRIX DECOMPOSITIONS AND APPLICATIONS
David Skillicorn Ashok N. Srivastava and Mehran Sahami
COMPUTATIONAL METHODS OF FEATURE BIOLOGICAL DATA MINING
SELECTION Jake Y. Chen and Stefano Lonardi
Huan Liu and Hiroshi Motoda
INFORMATION DISCOVERY ON ELECTRONIC
CONSTRAINED CLUSTERING: ADVANCES IN HEALTH RECORDS
ALGORITHMS, THEORY, AND APPLICATIONS Vagelis Hristidis
Sugato Basu, Ian Davidson, and Kiri L. Wagstaff
TEMPORAL DATA MINING
KNOWLEDGE DISCOVERY FOR Theophano Mitsa
COUNTERTERRORISM AND LAW ENFORCEMENT
David Skillicorn RELATIONAL DATA CLUSTERING: MODELS,
ALGORITHMS, AND APPLICATIONS
MULTIMEDIA DATA MINING: A SYSTEMATIC Bo Long, Zhongfei Zhang, and Philip S. Yu
INTRODUCTION TO CONCEPTS AND THEORY
Zhongfei Zhang and Ruofei Zhang KNOWLEDGE DISCOVERY FROM DATA STREAMS
João Gama
NEXT GENERATION OF DATA MINING
Hillol Kargupta, Jiawei Han, Philip S. Yu, STATISTICAL DATA MINING USING SAS
Rajeev Motwani, and Vipin Kumar APPLICATIONS, SECOND EDITION
George Fernandez
DATA MINING FOR DESIGN AND MARKETING
Yukio Ohsawa and Katsutoshi Yada INTRODUCTION TO PRIVACY-PRESERVING DATA
PUBLISHING: CONCEPTS AND TECHNIQUES
THE TOP TEN ALGORITHMS IN DATA MINING Benjamin C. M. Fung, Ke Wang, Ada Wai-Chee Fu,
Xindong Wu and Vipin Kumar and Philip S. Yu
GEOGRAPHIC DATA MINING AND HANDBOOK OF EDUCATIONAL DATA MINING
KNOWLEDGE DISCOVERY, SECOND EDITION Cristóbal Romero, Sebastian Ventura,
Harvey J. Miller and Jiawei Han Mykola Pechenizkiy, and Ryan S.J.d. Baker
Chapman & Hall/CRC
Data Mining and Knowledge Discovery Series
Handbook of
Educational Data Mining
Edited by
Cristóbal Romero, Sebastian Ventura,
Mykola Pechenizkiy, and Ryan S.J.d. Baker
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the
accuracy of the text or exercises in this book. This book’s use or discussion of MATLAB® software or related products
does not constitute endorsement or sponsorship by The MathWorks of a particular pedagogical approach or particular
use of the MATLAB® software.
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 uti-
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.
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, Ana, and my son, Cristóbal
Cristóbal Romero
To my wife, Inma, and my daughter, Marta
Sebastián Ventura
To my wife, Ekaterina, and my daughter, Aleksandra
Mykola Pechenizkiy
To my wife, Adriana, and my daughter, Maria
Ryan S. J. d. Baker
Contents
Preface...............................................................................................................................................xi
Editors..............................................................................................................................................xv
Contributors................................................................................................................................. xvii
1. Introduction..............................................................................................................................1
Cristóbal Romero, Sebastián Ventura, Mykola Pechenizkiy, and Ryan S. J. d. Baker
Part Iâ•… 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 Trčka, Mykola Pechenizkiy, and Wil van der Aalst
10. Modeling Hierarchy and Dependence among Task Responses in

Educational Data Mining...................................................................................................143
Brian W. Junker
vii
viii Contents
Part IIâ•… 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

Environments.......................................................................................................................231
Manolis Mavrikis, Sidney D’Mello, Kaska Porayska-Pomsta, Mihaela Cocea, and
Art Graesser
17. Measuring Correlation of Strong Symmetric Association Rules in

Educational Data..................................................................................................................245
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
Contents ix
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
Preface
The Purpose of This Book

The goal of this book is to provide an overview of the current state of knowledge of educa-
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.
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.
No one discipline has the necessary expertise to conduct EDM research. Thus, the com-
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
xi
xii Preface
are strong in statistical and computational techniques, but techniques and data are not suf-
ficient to advance a scientific domain; researchers with basic understanding of the teach-
ing and learning process are also required. Thus, education researchers and psychologists
are key participants in the EDM community.
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.
The first avenue is work on developing computational tools and techniques, determin-
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.
This work on extending and better understanding our computational toolkit is a neces-
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.
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
Preface xiii
teacher is a better choice. The first example is at the edge of what EDM is capable; the sec-
ond is, for now, beyond our capabilities.
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.
The field of EDM has grown substantially in the past five years, with the first work-
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.
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.
xiv Preface
MATLAB® is a registered trademark of The MathWorks, Inc. For product information,

please contact:
The MathWorks, Inc.
3 Apple Hill Drive
Natick, MA 01760-2098 USA
Tel: 508 647 7000
Fax: 508-647-7001
E-mail: [email protected]
Web: www.mathworks.com
Joseph E. Beck
Worcester Polytechnic Institute, Massachusetts
Editors
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-Â�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.
xv
xvi Editors
Dr. Ryan S. J. d. Baker is an assistant professor of psychol-

ogy 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
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.
Contributors
Wil van der Aalst Tiffany Barnes

Department of Mathematics and Department of Computer Science
Computer Science University of North Carolina at
Eindhoven University of Technology Charlotte
Eindhoven, the Netherlands Charlotte, North Carolina
Analía Amandi Joseph E. Beck

ISISTAN Research Institute Computer Science Department
Universidad Nacional del Centro de la Worcester Polytechnic Institute
Provincia de Buenos Aires Worcester, Massachusetts
Tandil, Argentina
and
Saleema Amershi Machine Learning Department

Department of Computer Science and Carnegie Mellon University
Engineering Pittsburgh, Pennsylvania
University of Washington
Seattle, Washington
Javier Bravo
Escuela Politécnica Superior
Brigham S. Anderson Universidad Autónoma de Madrid
School of Computer Science Madrid, Spain
Carnegie Mellon University
Pittsburgh, Pennsylvania
Winslow Burleson
Department of Computer Science and
Cláudia Antunes
Engineering
Department of Computer Science and Arizona State University
Engineering Tempe, Arizona
Instituto Superior Técnico
Lisbon, Portugal
Toon Calders
Ivon Arroyo Department of Mathematics and
Department of Computer Science Computer Science
University of Massachusetts Amherst Eindhoven University of Technology
Amherst, Massachusetts Eindhoven, the Netherlands
Ryan S. J. d. Baker Cristina Carmona

Department of Social Science and Policy Departamento de Lenguajes y Ciencias de
Studies la Computación
Worcester Polytechnic Institute Universidad de Málaga
Worcester, Massachusetts Málaga, Spain
xvii
xviii Contributors
Gladys Castillo Andrew Cuneo

Department of Mathematics Robotics Institute
Centre for Research on Optimization and Carnegie Mellon University
Control Pittsburgh, Pennsylvania
University of Aveiro
Aveiro, Portugal Kyle Cunningham
Human–Computer Interaction Institute
Carlos de Castro Carnegie Mellon University
Department of Computer Science and Pittsburgh, Pennsylvania
Numerical Analysis
University of Cordoba
Cordoba, Spain Sidney D’Mello
Institute for Intelligent Systems
Félix Castro The University of Memphis
Departament de Llenguatges i Sistemes Memphis, Tennessee
Informàtics
Universitat Politècnica de Catalunya Philippe Fournier-Viger
Barcelona, Spain Department of Computer Science
and University of Quebec in Montreal
Montreal, Quebec, Canada
Centro de Investigación en Tecnologías de
Información y Sistemas
Universidad Autónoma del Estado de Enrique García
Hidalgo Department of Computer Science and
Hidalgo, Mexico Numerical Analysis
Mihaela Cocea Cordoba, Spain
London Knowledge Lab
The University of London
Daniela Godoy
London, United Kingdom
ISISTAN Research Institute
and Universidad Nacional del Centro de la
School of Computing Provincia de Buenos Aires
Tandil, Argentina
National College of Ireland
Dublin, Ireland
Evandro Gouvea
Cristina Conati
European Media Laboratory GmbH
Department of Computer Science
Heidelberg, Germany
University of British Columbia
Vancouver, British Columbia, Canada and
David G. Cooper Robotics Institute

Department of Computer Science Carnegie Mellon University
University of Massachusetts Amherst Pittsburgh, Pennsylvania
Amherst, Massachusetts
Marvin Croy Art Graesser

Department of Philosophy Institute for Intelligent Systems
University of North Carolina at Charlotte The University of Memphis
Charlotte, North Carolina Memphis, Tennessee
Contributors xix
Wilhelmiina Hämäläinen Jihie Kim

Department of Computer Science Information Sciences Institute
University of Helsinki University of Southern California
Helsinki, Finland Marina del Rey, California
Cristina L. Heffernan
Kenneth R. Koedinger
Worcester Polytechnic Institute
Worcester, Massachusetts
Neil T. Heffernan
Department of Computer Science Irena Koprinska
Worcester Polytechnic Institute School of Information Technologies
Worcester, Massachusetts University of Sydney
Sydney, New South Wales, Australia
Cecily Heiner
Language Technologies Institute
Brett Leber
and Pittsburgh, Pennsylvania
Computer Science Department
University of Utah Tara M. Madhyastha
Salt Lake City, Utah Department of Psychology
Arnon Hershkovitz Seattle, Washington
Knowledge Technology Lab
School of Education David Masip
Tel Aviv University Department of Computer Science,
Tel Aviv, Israel
Multimedia and Telecommunications
Universitat Oberta de Catalunya
Earl Hunt
Barcelona, Spain
Department of Psychology
Seattle, Washington Manolis Mavrikis
London Knowledge Lab
Octavio Juarez The University of London
Robotics Institute London, United Kingdom
Riccardo Mazza
Brian W. Junker Faculty of Communication Sciences
Department of Statistics University of Lugano
Carnegie Mellon University Lugano, Switzerland
Pittsburgh, Pennsylvania and
Judy Kay Department of Innovative Technologies
School of Information Technologies University of Applied Sciences of Southern
University of Sydney Switzerland
Sydney, New South Wales, Australia Manno, Switzerland
xx Contributors
Gordon G. McCalla Àngela Nebot

Department of Computer Science Departament de Llenguatges i Sistemes
University of Saskatchewan Informàtics
Saskatoon, Saskatchewan, Canada Universitat Politècnica de Catalunya
Barcelona, Spain
Victor H. Menendez
Facultad de Matemáticas John C. Nesbit
Universidad Autónoma de Yucatán Faculty of Education
Merida, Mexico Simon Fraser University
Burnaby, British Columbia, Canada
Agathe Merceron
Media and Computer Science Department Engelbert Mephu Nguifo
Beuth University of Applied Sciences Department of Computer Sciences
Berlin, Germany Université Blaise-Pascal Clermont 2
Clermont-Ferrand, France
Eva Millán Roger Nkambou

Departamento de Lenguajes y Ciencias de Department of Computer Science
la Computación University of Quebec in Montreal
Universidad de Málaga Montreal, Quebec, Canada
Malaga, Spain
Alvaro Ortigosa
Escuela Politécnica Superior
Julià Minguillón
Universidad Autónoma de Madrid
Department of Computer Science,
Madrid, Spain
Multimedia and Telecommunications
Universitat Oberta de Catalunya
Zachary A. Pardos
Barcelona, Spain
Worcester Polytechnic Institute
Enric Mor Worcester, Massachusetts
Department of Computer Science,
Multimedia and Telecommunications Mykola Pechenizkiy
Universitat Oberta de Catalunya Department of Mathematics and
Barcelona, Spain Computer Science
Eindhoven University of Technology
Eindhoven, the Netherlands
Jack Mostow
Robotics Institute Kaska Porayska-Pomsta
Carnegie Mellon University London Knowledge Lab
Pittsburgh, Pennsylvania The University of London
London, United Kingdom
Rafi Nachmias
Knowledge Technology Lab Manuel E. Prieto
School of Education Escuela Superior de Informática
Tel Aviv University Universidad de Castilla-La Mancha
Tel Aviv, Israel Ciudad Real, Spain
Contributors xxi
Sujith Ravi Nikola Trčka

Information Sciences Institute Department of Mathematics and
University of Southern California Computer Science
Marina del Rey, California Eindhoven University of Technology
Eindhoven, the Netherlands
Cristóbal Romero
Alfredo Vellido
Numerical Analysis
Departament de Llenguatges i Sistemes
Informàtics
Cordoba, Spain
Universitat Politècnica de Catalunya
Barcelona, Spain
Richard Scheines
Department of Philosophy
Carnegie Mellon University Sebastián Ventura
Pittsburgh, Pennsylvania Department of Computer Science and
Numerical Analysis
Erin Shaw
Cordoba, Spain
Information Sciences Institute
University of Southern California
Marina del Rey, California César Vialardi
Facultad de Ingeniería de Sistemas
Judy Sheard Universidad de Lima
Faculty of Information Technology Lima, Peru
Monash University
Melbourne, Victoria, Australia
Mikko Vinni
Benjamin Shih School of Computing
University of Eastern Finland
Machine Learning Department
Joensuu, Finland
Stephan Weibelzahl
Alida Skogsholm School of Computing
Human–Computer Interaction Institute National College of Ireland
Carnegie Mellon University Dublin, Ireland
John Stamper Philip H. Winne

Human–Computer Interaction Institute Faculty of Education
Carnegie Mellon University Simon Fraser University
Pittsburgh, Pennsylvania Burnaby, British Columbia, Canada
Tiffany Y. Tang Beverly P. Woolf

Department of Computer Science Department of Computer Science
Konkuk University University of Massachusetts Amherst
Chungju-si, South Korea Amherst, Massachusetts
xxii Contributors
Yabo Xu Alfredo Zapata

School of Computing Science Facultad de Educación
Simon Fraser University Universidad Autónoma de Yucatán
Burnaby, British Columbia, Canada Merida, Mexico
Kalina Yacef Mingming Zhou

School of Information Technologies Faculty of Education
University of Sydney Simon Fraser University
Sydney, New South Wales, Australia Burnaby, British Columbia, Canada
Amelia Zafra
Numerical Analysis
Cordoba, Spain
1
Introduction
Cristóbal Romero, Sebastián Ventura, Mykola Pechenizkiy, and Ryan S. J. d. Baker
Contents
1.1 Background.............................................................................................................................. 1
1.2 Educational Applications....................................................................................................... 3
1.3 Objectives, Content, and How to Read This Book............................................................. 4
References..........................................................................................................................................5
1.1â•‡ Background
In the last years, researchers from a variety of disciplines (including computer science,
statistics, data mining, and education) have begun to investigate how data mining can
improve education and facilitate education research. Educational data mining (EDM) is
increasingly recognized as an emerging discipline [10]. EDM focuses on the development
of methods for exploring the unique types of data that come from an educational context.
These data come from several sources, including data from traditional face-to-face class-
room environments, educational software, online courseware, and summative/high-stakes
tests. These sources increasingly provide vast amounts of data, which can be analyzed to
easily address questions that were not previously feasible, involving differences between
student populations, or involving uncommon student behaviors. EDM is contributing to
education and education research in a multitude of ways, as can be seen from the diver-
sity of educational problems considered in the following chapters of this volume. EDM’s
contributions have influenced thinking on pedagogy and learning, and have promoted
the improvement of educational software, improving software’s capacity to individualize
students’ learning experiences. As EDM matures as a research area, it has produced a con-
ference series (The International Conference on Educational Data Mining—as of 2010, in
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:
1
2 Handbook of Educational Data Mining
• 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
Introduction 3
Educational systems
(traditional classrooms, e-learning
Use, interact with,
systems, LMSs, web-based
participate in, design,
adaptive systems, intelligent
plan, build and maintain
tutoring systems, questionnaires
and quizzes)
Users
(students, learners, Provide, store:
instructors, Course information, contents,
teachers, course academic data, grades,
administrators, academic student usage and interaction data
researchers, school district
officials) Data mining techniques
(statistics, visualization, clustering,
Model learners and learning, classification, association rule
communicate findings, make mining, sequence mining, text
recommendations mining)
FIGURE 1.1
Applying data mining to the design of educational systems.
pedagogical basis for decisions when designing or modifying an environment’s pedagogi-

cal approach.
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.
1.2â•‡ 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.
Chapters 2 through 4, 9, 12, 22, 24, and 28 discuss methods and case studies for this
category of application.
• 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.
1.3â•‡ 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
Introduction 5
courses and primary and secondary schools. For instance, 6% of U.S. high schools now use
Cognitive Tutor software for mathematics learning (cf. [6]). As these environments become
more widespread, ever-larger collections of data have been obtained by educational data
repositories. A case study on one of the largest of these repositories is given in the chapter
on the PSLC DataShop by Koedinger and colleagues.
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.
References
1. Arruabarrena, R., Pérez, T. A., López-Cuadrado, J., and Vadillo, J. G. J. (2002). On evaluating
adaptive systems for education. In International Conference on Adaptive Hypermedia and Adaptive
Web-Based Systems, Málaga, Spain, pp. 363–367.
2. Baker, R.S.J.d. and Yacef, K. (2009). The state of educational data mining in 2009: A review and
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),
267–287.
4. Hirschman, L., Park, J.C., Tsujii, J., Wong, W., and Wu, C.H. (2002). Accomplishments and chal-
lenges in literature data mining for biology. Bioinformatics, 18(12), 1553–1561.
5. Ingram, A. (1999). Using web server logs in evaluating instructional web sites. Journal of
Educational Technology Systems, 28(2), 137–157.
6. Koedinger, K. and Corbett, A. (2006). Cognitive tutors: Technology bringing learning science to
the classroom. In K. Sawyer (Ed.), The Cambridge Handbook of the Learning Sciences. Cambridge,
U.K.: Cambridge University Press, pp. 61–78.
7. Lewis, M. (2004). Moneyball: The Art of Winning an Unfair Game. New York: Norton.
8. Li, J. and Zaïane, 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.
11. Srivastava, J., Cooley, R., Deshpande, M., and Tan, P. (2000). Web usage mining: Discovery and
applications of usage patterns from web data. SIGKDD Explorations, 1(2), 12–23.
12. Shaw, M.J., Subramanian, C., Tan, G.W., and Welge, M.E. (2001). Knowledge management and
data management for marketing. Decision Support Systems, 31(1), 127–137.
Part I
Basic Techniques, Surveys

and Tutorials
2
Visualization in Educational Environments
Riccardo Mazza
Contents
2.1 Introduction.............................................................................................................................9
2.2 What Is Information Visualization?................................................................................... 10
2.2.1 Visual Representations............................................................................................ 10
2.2.2 Interaction.................................................................................................................. 11
2.2.3 Abstract Data............................................................................................................. 11
2.2.4 Cognitive Amplification.......................................................................................... 12
2.3 Design Principles.................................................................................................................. 13
2.3.1 Spatial Clarity............................................................................................................ 14
2.3.2 Graphical Excellence................................................................................................. 14
2.4 Visualizations in Educational Software............................................................................ 16
2.4.1 Visualizations of User Models................................................................................ 16
2.4.1.1 UM/QV........................................................................................................ 16
2.4.1.2 ViSMod........................................................................................................ 17
2.4.1.3 E-KERMIT................................................................................................... 18
2.4.2 Visualizations of Online Communications.......................................................... 19
2.4.2.1 Simuligne.................................................................................................... 19
2.4.2.2 PeopleGarden............................................................................................. 20
2.4.3 Visualizations of Student-Tracking Data............................................................... 20
2.5 Conclusions............................................................................................................................ 24
References........................................................................................................................................ 25
2.1â•‡ 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.
9
2.2â•‡ What Is Information Visualization?

Visualization, which may be defined as “the display of data with the aim of maximiz-
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.
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.
2.2.1â•‡ Visual Representations

There are many situations in the real world where we try to understand some phenom-
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 250â•›ms or less) are called
pre-attentive, since they occur without the intervention of consciousness [20].
According to Ware [20], the graphical proprieties that are pre-
attentively processed can be grouped into four basic categories: 320
color, form, movement, and spatial position. Managing properly
the elements that are “pre-attentively” processed can make a 260
difference in a user interface and is fundamental for the genera- 380
tion of good user interfaces and graphics.
Let us try to explain these concepts with a practical example. 280
In Figure 2.1 are represented a list of numbers together with 420
bars whose length is proportional to the number on the left.
Suppose we have to find the maximum and the minimum 400
between the numbers. If we don’t have the bars on the right, we FIGURE 2.1
need to read and understand each value, and take into account Comparing perception of
the maximum and the minimum value we read until we reach lines with numbers.
* A Dictionary of Computing. Oxford University Press, 1996. Oxford Reference Online. Oxford University Press.
Visualization in Educational Environments 11
the end. This can be defined as a scrutiny task, because it is a conscious operation that
involves memory, semantics, and symbolism.
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.
Graphical representations are often associated with the term “visualization” (or “visu-
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:
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.
2.2.2â•‡ 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.
2.2.3â•‡ 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.
‡ Merriam-Webster Online Dictionary. http://www.webster.com
earth), and data that is more abstract in nature (e.g., the stock
market fluctuations). The former is known as scientific visu-
alization, and the latter as IV [4,16,19].
Scientific visualization was developed in response to the
needs of scientists and engineers to view experimental or
phenomenal data in graphical formats (an example is given
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
with some examples of abstract data and physical data.
However, we ought to say that this distinction is not strict,
and sometimes abstract and physical data are combined in a FIGURE 2.2
single representation. For instance, the results from the last Example of scientific visualiza-
Swiss federal referendum on changing the Swiss law on asy- tion: The ozone hole the South
lum can be considered a sort of abstract data if the goal of the Pole on September 22, 2004.
graphical representation is to highlight the preference (yes (Image from the NASA Goddard
Space Center archives and repro-
or no) with respect to the social status, age, sex, etc. of the duced with permission.)
voter. But if we want to highlight the percentage that the ref-
erendum got in each town, a mapping with the geographical location might be helpful to
see how the linguistic regions, cantons, and the proximity with the border influenced the
choice of the electorate (see Figure 2.3).
2.2.4â•‡ Cognitive Amplification

Graphics aid thinking and reasoning in several ways. For example, let us take a multi-
plication (a typical mental activity), e.g., 27â•›×â•›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:
2 7 ×
4 2
5 4
1 0 8 —
1 1 3 4
This example shows how visual and manipulative use of the external representations and
processing amplifies cognitive performance. Graphics use the visual representations that
help to amplify cognition. They convey information to our minds that allows us to search
TABLE 2.1
Some Examples of Abstract Data and
Physical Data
Abstract Data Physical Data
Names Data gathered from
instruments
Grades Simulations of wind flow
News or stories Geographical locations
Jobs Molecular structure
Änderung des Asylgesetzes

Modification de la loi sur l΄asile
Ja-Stimmenanteil nach Bezirken
proportion de «oui» par district
≥70% SH
65–69.9% BS
60–64.9% TG
55–59.9%
AG
50–54.9% ZH
45–49.9% BL
JU AR
SG
SO AI
ZG
LU
Provisorische Ergebnisse SZ
Résultats provisoires NE GL
NW
BE UR
OW GR
FR
VD
GE
TI
VS 0 25 50 km
Schweiz/Suisse
Stimmbeteiligung/participation: 48.4%
Abstimmung vom 24. September 2006
Ja-Stimmenanteil/proportion de «oui»: 67.8%
Votation du 24 Septembre 2006
Abst.–Nr./n ° vot.: 525
Schweizerische Eidgenossenschaft Eidgenössisches Departement des Innern EDI Quelle: Abstimmungsstatistik, BFS
Confédération suisse Département fédéral de l΄intérieur DFI Source: Statistique des votations, OFS
Confederazione svizzera Bundesamt für Statistik BFS © BFS, ThemaKart, Neuenburg 2006/K17.A525.R_bz
Confederaziun svizra Office fédéral de la statistique OFS © 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.)
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

2. Reducing the search for information
3. Using visual representations to enhance the detection of patterns
4. Enabling perceptual inference operations
5. Using perceptual mechanisms for monitoring
6. Encoding information in a manipulable medium
2.3â•‡ Design Principles

The basic principle of IV is to present data in form of graphical representations and use
human perceptual abilities for their interpretation. “A picture is worth a thousand words”
is a well-known old adage that everybody knows. But why (and in which situations)
graphical representations are effective?
2.3.1â•‡ Spatial Clarity

Graphical representations may facilitate the way we present and understand large Â�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 enabled by using location to group information about a single ele-
ment, 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.
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.
2.3.2â•‡ 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]
• Show the data.

• Induce the viewer to think about the substance, not the methodology.
• Avoid distorting what the data says.
• Present many data in a small space.
• Give different perspectives on the data—from a broad overview to a fine structure.
• Serve a reasonably clear purpose: description, exploration, tabulation, or decoration.
• Be closely integrated with the statistical and verbal descriptions of the data set.
• Encourage inferential processes, such as comparing different pieces of data.
• Give different perspectives on the data—from a broad overview to a fine structure.
La Manuel de Reyes
Moraleja Falla Católicos
10 Hospital del Norte
La Granja Marqués Baunatal
Ronda de la Comunicación de la
Las Tablas Valdavia
Montecarmelo
Tres Olivos
Herrera Oria 9 Fuencarral Pinar de chamartin Parque de Aeropuerto T4
Pitis 1 Manoteras
Barrio del Pilar Begoña Santa María 8
7 4
Ventilla San lorenzo
Arroyo del Fresno Bambú Hortaleza Barajas
Chamartín Mar de
Lacoma Pío XII Pinar del Rey Aeropuerto T1-T2-T3
Valdeacederas Cristal
Avenida de la Illustración
Peñagrande Duque de Campo de las
Tetuán Plaza de Colombia
Antonio Machado Castilla
Pastrana Canillas Naciones
Valdezarza Estrecho Cuzco Esperanza Alameda
Santiago Concha de Osuna
Francos Rodriguez Alvarado Arturo Soria 5
Bernabéu Espina
Cuatro Nuevos República Cruz del El Capricho
Guzmán el Bueno Caminos Ministerios Argentina Avenida de la Paz Canillejas
2 8 Rayo
Metropolitano Alfonso XIII Torre Arias
Ciudad Rios Rosas 6 Prosperidad Suanzes
Universitária Islas Alonso Gregorio Pque. de las Ciudad Lineal
Filipinas Cano Marañón Cartagena Avenidas Barrio de la
Canal
3 Moncloa Concepción
Avda. de América Diego de León El
Quevedo Iglesia
Argüelles San Bernardo Rubén Carmen
4 Núñez Ventas Pueblo Nuevo
Dario Quintana
Bilbao de Balboa
Ventura Lista 2 La Elipa
Rodriguez Noviciado Colón Velázquez Ascao
Pza. de España Garcia Noblejas
Tribunal Alonso Serrano Goya Manuel Becerra Simancas
Martínez San Blas
Santo Gran Príncipe de Vergara Las Musas
Príncipe Pío R O΄Donnell
domingo Via Chueca
Retiro Estadio Olimpico
Lago Callao Ibiza
Sevilla Barrio del Puerto
6 Sainz de Baranda Coslada Central
Batán Pta. del Banco de La Rambla
Angel Ópera R España San Fernando
Sol
5 Casa de Tirso de Molina Estrella Jarama
Alto de La
Campo Antón Martin Vinateros
Extremadura Latina Conde de 7
Campamento Lucero Pta. de Lavapiés Atocha Artilleros Henares
6 Casal
Empalme Toledo Atocha Renfe Pavones
Aluche Laguna Valdebernardo
Acacias Embajadores
Carpetana Menéndez Vicálvaro
Pirámides Pacífico
Eugenia Pelayo San Cipriano
de Montijo Urgel Palos de la Frontera Puente de Vallecas
Marqués Méndez
Delicias Nueva Numancia Puerta de Arganda
Carabanchel Vista Oporto de Vadillo Álvardo
Portazgo
Alegre Arganzuela-
Usera Buenos Aires Rivas Urbanizaciones
Colonia Jardin Planetario
Opañel 11 Alto del Arenal
Legazpi
Plaza Miguel Hernández Rivas Vaciamadrid
Aviación Española Eliptica Almendrales
Abrantes
Hospital 12 de Octubre Sierra de Guadalupe
Pan Bendito San Fermin-Orcasur Villa de Vallecas La Poveda
Cuatro Vientos
San Francisco Ciudad de los Ángeles Congosto
Carabanchel Alto Villaverde Bajo-Cruce La Gavia 9 Arganda del Rey
Joaquin Vilumbrales San Cristóbal Las Suertes
11 La Peseta
3 Villaverde Alto 1 Valdecarros
Puerta Leganés Hospital Casa del Julián
del Sur 12 San Nicasio Central Severo Ochoa Reloj Besteiro El Carrascal El Bercial
10
Metro de Madrid Los Espartales
Parque Lisboa El Casar
© 2007 Designed and drawn by Matthew McLauchlin, http://www.metrodemontreal.com/ 12
Alcorcón Central This version released under Creative Commons Share-Alike Attribution Licence (CC-SA-BY 2.5) Juan de la Cierva
Parque Oeste http://creativecommons.org/licenses/by-sa/2.5/ Getafe Central
Universidad Rey Juan Carlos Not affiliated with, released by, or approved of by the Alonso de Mendoza
Móstoles Central Consorcio Regional de Transportes de Madrid (http://www.metromadrid.es/) Conservatorio
Pradillo Arroyo Culebro
12
Hospital Manuela Loranca Hospital de Parque Fuenlabrada Parque de los
de Móstoles Malasaña Fuenlabrada Europa Central Estados
FIGURE 2.4
Map of the Madrid metro system. (Images licensed under Creative Commons Share-Alike.)
Following these principles is the key to build what Tufte calls the graphical excellence, and
it consists in “giving the viewer the greatest number of ideas in the shortest time with the
least ink in the smallest space” [18].
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 com-

puter screen or on a paper are presented in a 2D surface.
2. Sometimes we need to represent a huge dataset, while the number of data view-
able on a computer screen or on a paper is limited.
3. Data may vary during the time, while graphics are static.
4. Humans have remarkable abilities to select, manipulate, and rearrange data, so
the graphical representations should provide users with these features.
2.4â•‡ 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.
2.4.1â•‡ Visualizations of User Models

A user model is a representation of a set of beliefs about the user, particularly their knowl-
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].
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
Select node Label Label

Quit No
to fold all leaves
overview labels
or unfold nodes only
Minimal
quit_k
Basics Useful quit_b
default_size_k
non_cmd_b
undo_k
load_new_k
gotoline_k
command_window
more_useful set_fname_k
Sam write_k
Editors Mouse
Other
command_window
Root exch_k
search_k
very_useful
mouse xerox_k
Powerful
mostly_useless
emacs
vi
c_c
Programming pascal_c
languages
lisp_c
fortran_c
typing_ok_c
user_info
input wpm_info
FIGURE 2.5
The QV tool showing a user model. (Image courtesy of Judy Kay.)
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
FIGURE 2.6
A screenshot of ViSMod showing a fragment of a Bayesian student model in the area of biology cell. (Image
courtesy of Diego Zapata-Rivera.)
Bayesian student model and various visualization techniques such as color, size proximity
link thickness, and animation to represent concepts such as marginal probability, changes
in probability, probability propagation, and cause–effect relationships (Figure 2.6). One
interesting aspect of this model is that the overall belief of a student knowing a particular
concept is captured taking into account the students’ opinion, the instructors’ opinion, and
the influence of social aspects of learning on each concept. By using VisMod, it is possible
to inspect complex networks by focusing on a particular segment (e.g., zooming or scroll-
ing) and using animations to represent how probability propagation occurs in a simple
network in which several causes affect a single node.
2.4.1.3 E-KERMIT
KERMIT (Knowledge-based Entity Relationship Modelling Intelligent Tutor) [17] is a
knowledge-based intelligent tutoring system aimed at teaching conceptual database
design for university level students. KERMIT teaches the basic entity-relationship (ER)
database modeling by presenting to the student the requirements for a database, and the
student has to design an ER diagram for it. E-KERMIT is an extension of KERMIT devel-
oped by Hartley and Mitrovic [3] with an open student model. In E-KERMIT the student
may examine with a dedicated interface the global view of the student model (see Figure
2.7). The course domain is divided in categories, representing the processes and concepts
in ER modeling. In the representation of the open student model concepts of the domain
are mapped with histograms. The histogram shows how much of the concrete part of the
domain the student knows correctly (in black) or incorrectly (in gray) and the percentage
of covered on the concepts of the category. For instance, the example shows that the stu-
dent covered 32% of the concepts of the category Type, and has scored 23% out of a possible
32% on this category. This means that the student’s performance on category type so far is
77% (23/320â•›×â•›100).
– 37% (51%) Knowledge
+ 66% (100%) Notation
– 32% (44%) Attribute identification
+ 23% (32%) Type
– 35% (47%) Structure

17% (33%) Simple
45% (55%) Composite
+ 23% (32%) Num values
70% (100%) Key

89% (100%) Partial key
FIGURE 2.7
The main view of a student’s progress in E-KERMIT. Progress bars indicate how much the student compre-
hends each category of the domain. (Image reproduced with permission of Tanja Mitrovic.)
2.4.2â•‡ Visualizations of Online Communications

One important aspect that should be considered in distance education is the social one.
There are various types of interactions that may occur in distance learning, such as inter-
actions between individual students and interactions between students and the instruc-
tors. In the Internet, the main tools that engage students and instructors in communicative
activities are discussion forums, e-mail, and chat. Some works attempted to visualize
the communications exchanged in educational settings to discover patterns, roles, and
engagement of students in social activities.
2.4.2.1 Simuligne
Simuligne [12] is a research project that uses social network analysis [14] to monitor
group communications in distance learning in order to help instructors detect collabora-
tion problems or slowdown of group interactions. Social network analysis is a research
field that “characterize the group’s structure and, in particular, the influence of each of
the members on that group, reasoning on the relationship that can be observed in that
group” (Reffay and Chanier [12], p. 343). It provides both a graphical and a mathematical
analysis of interactions between individuals. The graphical version can be represented
with a network, where the nodes in the network are the individuals and groups while
the links show relationships or flows between the nodes. The social network analysis can
help to determine the prominence of a student respect to others, and other social net-
work researcher measures, such as the cohesion factor between students. The cohesion
is a statistical measure that represents how much the individuals socialize in a group
that shares goals and values. Reffay and Chanier applied this theory to a list of e-mails
exchanged in a class of distance learners. Figure 2.8 illustrates the graphical representa-
tion of the e-mail graph for each learning group. We can see for instance that there is
no communication with Gl2 and Gl3, or the central role of the tutor in the discussions
(node Gt).
Gl1
28 25
2 4 Gl
2 20 2 19 12 17 4 2 9 3 2 1
Gl2 11 13 Gl3 Gl10 14 12 Gl5 Gl9
1 1 2 9 4 2 13
Gn2 22 Gl6 24
3 3 8 10
3 3 Gl4 3
4 5
Gn1
FIGURE 2.8
The communication graph of the e-mail exchanged within groups in Simuligne. (Image courtesy of Christophe
Reffay.)
2.4.2.2 PeopleGarden
PeopleGarden [21] uses a flower and garden metaphor to visualize participations on a
message board. The message board is visualized as a garden full of flowers. Each flower
represents one individual. The height of flower denotes amount of time a user has been
at the board and its petals his postings. Initial postings are shown in red, replies in blue.
An example is represented in Figure 2.9. The figure can help the instructor of a course to
quickly grasp the underlying situation, such as a single dominant member in discussion
on the left or a group with many members at different level of participation on the right.
2.4.3â•‡ Visualizations of Student-Tracking Data

Many software environments have been produced in the latest years to facilitate the adop-
tion of distance learning through the Web. Thanks to the diffusion and the popularity of
the Web, many software environments take advantage of the client–server communication
on the Internet to support distance learning. Environments called course management
systems (CMSs) have been developed for this purpose. Using CMSs, instructors can dis-
tribute information to students, produce content material, prepare assignments and tests,
engage in discussions, and manage distance classes.
One of the most common problems with CMSs is the difficulty for the instructor to
verify what students are doing: are the students attending the course, reading materials,
and performing exercises? CMSs accumulate tracking data in form of large log in a dis-
tance course, and it can be used to monitor the students’ activities. This data is accessible
FIGURE 2.9
The PeopleGarden visual representations of participation on a message board. (Image by Rebecca Xiong and
Judith Donath, © 1999 MIT media lab.)
to the instructor of the course, but it is commonly presented in the format of a textual log
file, which is inappropriate for the instructor’s needs [8]. To this end, since the log data is
collected in a format that is suitable to be analyzed with IV techniques and tools, a number
of approaches have been proposed to graphically represent the tracking data generated
by a CMS.
Recently, a number of researches that exploits graphical representations to analyze the
student tracking have been proposed. ViSION [15] is a tool that was implemented to dis-
play student interactions with a courseware website designed to assist students with their
group project work. CourseVis [10] is another application that exploits graphical repre-
sentations to analyze the student tracking data. CourseVis is a visual student tracking
tool that transforms tracking data from a CMS into graphical representations that can be
explored and manipulated by course instructors to examine social, cognitive, and behav-
ioral aspects of distance students. CourseVis was started from a systematic investigation
aimed to find out what information about distance students the instructors need when
they run courses with a CMS, as well as to identify possible ways to help instructors
acquire this information. This investigation was conducted with a survey, and the results
were used to draw the requirements and to inform the design of the graphical representa-
tions. One of the (several) graphical representations produced by CourseVis is reported in
Figure 2.10.
This comprehensive image represents in a single view the behaviors of a specific stu-
dent in an online course. It takes advantage of single-axis composition method (multiple
variables share an axis and are aligned using that axis) for presenting large number of
variables in a 2D metric space. With a common x-axis mapping the dates of the course,
a number of variables are represented. The information represented here are namely the
student’s access to the content pages (ordered by topics of the course), the global access
to the course (content pages, quiz, discussion, etc.), a progress with the schedule of the
course, messages (posted, read, follow-ups), and the submission of quizzes and assign-
ments. For a detailed description of CourseVis, see Mazza and Dimitrova [10].
22
Summary of student’s behaviors from 2002-01-15 to 2002-04-11
Student: Francesco
Variable
Threads
String
Program structure
Package
Overriding
Overloading
Object serialization
Object
Method
Interface
Inner class
Access to Inheritance
I/O streams
content pages File
Exception
by topics Data file
Control flow
Constructor
Class variable
Class method
Class
Basic concepts
Array
Argument
Applet
Access level
Abstract method
Abstract class
Graphic libraries
AWT structure
AWT contexts
AWT components
AWT events
––––
Global accesses (hits)

to the course
Handbook of Educational Data Mining

Progress with
the course schedule
Original posts
Messages Follow ups
Articles read
Q Q Q Q Q Q QQ
Quiz (Q) and assignment (A) A
submissions
2002–01–15
2002–01–16
2002–01–17
2002–01–18
2002–01–19
2002–01–20
2002–01–21
2002–01–22
2002–01–23
2002–01–24
2002–01–25
2002–01–26
2002–01–27
2002–01–28
2002–01–29
2002–01–30
2002–01–31
2002–02–01
2002–02–02
2002–02–03
2002–02–04
2002–02–05
2002–02–06
2002–02–07
2002–02–08
2002–02–09
2002–02–10
2002–02–11
2002–02–12
2002–02–13
2002–02–14
2002–02–15
2002–02–16
2002–02–17
2002–02–18
2002–02–19
2002–02–20
2002–02–21
2002–02–22
2002–02–23
2002–02–24
2002–02–25
2002–02–26
2002–02–27
2002–02–28
2002–03–01
2002–03–02
2002–03–03
2002–03–04
2002–03–05
2002–03–06
2002–03–07
2002–03–08
2002–03–09
2002–03–10
2002–03–11
2002–03–12
2002–03–13
2002–03–14
2002–03–15
2002–03–16
2002–03–17
2002–03–18
2002–03–19
2002–03–20
2002–03–21
2002–03–22
2002–03–23
2002–03–24
2002–03–25
2002–03–26
2002–03–27
2002–03–28
2002–03–29
2002–03–30
2002–03–31
2002–04–01
2002–04–02
2002–04–03
2002–04–04
2002–04–05
2002–04–06
2002–04–07
2002–04–08
2002–04–09
2002–04–10
2002–04–11
FIGURE 2.10
Graphical representation of student’s behaviors in CourseVis. (Image from Mazza, R. and Dimitrova, V., Generation of graphical representations of student tracking
data in course management systems, in Proceedings of the 9th International Conference on Information Visualisation, London, U.K., July 6–8, 2005. © 2005 IEEE.)
Exploring the Variety of Random
Documents with Different Content
about two minutes, and was just crowing to myself when everything
seemed to give way at once, and the ice flew up and hit all my
knees and elbows, and there I was in a heap, with my skates locked
together as if they were a padlock. Rob sorted me out, and tried not
to laugh, till I told him to go ahead, and then he just roared! He said
if I’d only been lighted, I’d have made such a gorgeous pin-wheel!
“Perhaps you’ll think I’d had enough—I thought I had then myself,
but just before we started for home I believed I really had got the
hang of it this time, so I let go again. I struck out all right, and went
ahead for two or three yards, and Rob and Pep had just begun to
clap their hands and hurrah when before I knew what had happened
I was sure I felt my backbone coming out of the top of my head,
and there I was again, sitting down as flat as a pancake, and feeling
a good deal flatter! I didn’t try any more after that, but just took off
my skates and came home.”
Mrs. Leslie could not help smiling at this graphic account of
Johnny’s first attempt at skating, but when she tucked him up and
gave him his last kiss, she said,—
“Johnny, do you know of what your adventures to-day have made
me think? A verse in the Bible—‘Let him that thinketh he standeth
take heed lest he fall.’ Nearly all our falls come from being very sure
we can stand, and from refusing the offered help.”
“Pep didn’t fall once,” said Johnny, thoughtfully, “though it was his
first skate, too, and he’s younger than I am. Yes, I see what you
mean, mamma, and I hope I’ll remember it at the right time—but
I’m so apt not to remember till afterward!”
“That is why we are taught to ask that God’s grace ‘may always
prevent’—that is, go before to smooth the way—‘and follow us,’”
replied his mother, as she stooped to give him another last kiss.
Johnny applied his lesson to his next attempt at skating, and came
home triumphant, saying,—
“We didn’t fall once, mamma, either of us, and Rob let us go a
little way alone, but he skated backward, just in front of us, and
caught us every time we staggered much.”
But in two weeks, during which time the skating remained good,
Rob’s pupils ventured fearlessly all about the pond, without a helping
hand, and had even begun to try to cut letters and figures—though
not, it must be admitted, with any great amount of success. Mrs.
Leslie declared that she must see some of the wonderful
performances of which she heard so much, so one bright afternoon,
when the mildness of the air threatened to spoil their fun before
long, she wrapped Tiny and Polly warmly up, hired Mr. Chipman’s
safest horse and best wagon, and drove in state to the pond.
The boys were delighted, and did their best, but of course, in his
eagerness to excel himself, Johnny managed to fall once or twice,
and Rob was obliged to testify that this was now quite unusual.
Then they begged for Polly—Tiny had been allowed to leave the
wagon when it first arrived, and was successfully and joyfully sliding.
“Oh, do let us have Polly, if it’s just for five minutes, mamma!” said
Johnny, eagerly. “We’ll take off our skates and give her a slide. It’s
first-rate sliding, here by the bank, and it’s quite safe.”
So Miss Polly, chuckling with delight, was lifted from the wagon,
while Johnny and Pep pulled off their skates, but she was a little
frightened when she felt the slippery ice under her feet, and “hung
down like a rag doll,” as Johnny said, instead of putting herself in
sliding position.
THE SKATING LESSON.
“Stand up straight, Polly, and put your feet down flat, so,” said
Johnny, as Polly slid helplessly along on the backs of her heels,
resting all her little weight confidingly upon the boys. And, after two
or three earnest explanations from Johnny and Pep, she suddenly
seemed to understand; she stiffened up, grasped a hand on each
side, and went off in such style that the boys had almost to run to
keep up with her, and she obeyed her mother’s call very unwillingly.
“Wasn’t it fun to see her little face, though!” said Johnny, as he
and Pep walked home, having declined the proffered drive for the
sake of a little more skating. “I think she thought something had
made her feet slippery, all of a sudden—she’d never been on ice
before.”
The thaw came very soon after this, as thaws will come, even
when people have new steel skates, but happily, there are always
tops and marbles, and, as some wise person has remarked, “When
one door shuts, another opens.”
Johnny did not play marbles “for keeps”;
his father had explained to him that taking
anything without giving a fair return for it is
dishonesty, and as he quite understood this,
he had no desire to “win” marbles from
boys who could not shoot so well as he
could, but he enjoyed playing fully as much
as anybody did, and found the game
exciting enough when played merely for the
hope of victory.
It was in the midst of a very even game that the school bell rang
one morning. Johnny and one other boy were the champions; the
rest had “gone out.” They lingered for one more shot—two more—
then just a third to finish the game, and then, as they hurried into
the schoolroom, they found that the roll had been called, and they
were marked late.
Johnny had intended to take one more look at his history lesson,
but there was no time. He was sure of it all, except two or three
dates, and of course, one of those dates came to him—or rather,
didn’t come; it was the question that came. The next boy gave the
answer, and Johnny’s history lesson for the first time that term, was
marked “Imperfect.”
This vexed him so, that he gave only a small half of his mind to
his mental arithmetic, and he lost his place in the class,—lost it to a
boy who was almost certain to keep it, too.
Thinking of this misfortune, he dropped a penful of ink on his
spotless new copy-book, and, although he promptly licked it off, an
ugly smear remained, and the writing teacher reproved him for
untidiness. So he was very glad when two o’clock struck, and he
could go home and tell his mournful story, for he had an
uncomfortable feeling, under the injured one, that the real,
responsible cause of his misfortunes was one Johnny Leslie.
When his mother had heard it all with much sympathy, she paused
a moment, and then repeated these words,—
“‘That they who do lean only upon the hope of Thy Heavenly
grace, may evermore be defended by Thy mighty power.’”
A sudden light came into Johnny’s face, and he exclaimed,—
“That was it, mamma dear! I wasn’t leaning on it at all, and of
course, I went down! I know all about it now. I didn’t get up when
you called me the first time, and I said my prayers in a hurry, just as
if they were the multiplication table, and I didn’t wait to read the
verse in my little book—I meant to do it after breakfast, but the
marbles rattled in my pocket, and I forgot all about it, I was in such
a hurry to have a game before school. And I wouldn’t stop to think,
when the bell rang, except a sort of make-believe think that a
minute more would not make me too late to answer to my name,
and so I lost the chance to go over those dates. And the question I
missed in mental arithmetic was a mean little easy thing, if I’d had
my wits about me, but I was worrying about the history, and I made
that dreadful blot because I was thinking of both, and did not look,
and dug my pen down to the bottom of the inkstand. It’s just like
‘The House that Jack built.’”
“Yes,” said his mother, “I don’t think anything, the
smallest thing, stands quite alone; it is fast to
something else that it pulls after it, so we must keep
a sharp lookout for the first things. We can’t rub out
this bad day—it is like the blot on your copy book;
you will keep seeing the mark, even if you don’t
make another. But then, you can use the mark, with
the dear Saviour’s help, to keep you from making
another. To-morrow will be another day. You know
Tiny and you like Tennyson’s ‘Bugle Song’ so much,
here is something else he said,—
‘Men may rise on stepping-stones

Of their dead selves, to higher things.’
So to-morrow you must stand on this thoughtless, careless Johnny,

who forgets what he ought to remember, and be the Johnny you can
be, if you ‘lean only on the hope’ of that Heavenly grace which God
gives to His faithful children.”
It was an humble, but bright and hopeful Johnny who sprang up
at the first call the next morning, and started for school, with fresh
courage and resolution.
Try not to be defeated, little soldier, but, if defeats come, do you
too try to make them stepping-stones to victory.
CHAPTER VII.
THE EXTRA HORSE.
ohnny did not have a great deal of time for

thinking. It is difficult to think when one is
running, or jumping, or hammering, or
shouting, and still more difficult when one is
asleep! He often intended to “take a think”
about something that bothered him, after he
was in bed, and before he went to sleep, but
somehow, no matter how wide awake he
supposed he was before he began thinking,
he always found, before he had finished, that it was next morning,
and time to get up.
But he actually walked all the way home from school, one day,
without shouting once at anybody; he came and sat down in the
sewing-room, after he had put his books away, and was so quiet for
five minutes that his mother was just going to ask him if his head
ached, when he suddenly asked her,—
“Mamma, would you object to my keeping a peanut-stand—out of
school hours, you know, I mean?”
“Not at all,” replied Mrs. Leslie, “if you were obliged to earn your
living at once, and that were the only way in which you could
possibly do it. But papa and I are both anxious that you should earn
your living in a way which will help as many people as possible to
earn theirs, and if you were to set up a peanut-stand now, while you
are trying to learn a better way, I am afraid it would hinder our plans
for you.”
Johnny’s eyes had sparkled when his mother began with “Not at
all,” and now he looked a good deal disappointed.
“Yes, mamma,” he said, meekly, “I see that’s your side of it, but
may I just tell you my side?”
“Of course you may!” said Mrs. Leslie, smiling, and stopping her
sewing long enough to give him a hug and kiss. “I always like to
hear your side, even if I can’t agree with it, and I know you trust me
enough to come over to my side, even when you can’t see why.”
“It would be queer if I didn’t, mamma,” he said, drawing his stool
closer, and resting his arms on her knees, “you’ve come out right so
often when I was pretty sure you wouldn’t, you know. Now, its just
this way—I know you and papa aren’t rich, and I know I oughtn’t to
ask you for any more money than you give me now, but I do want
more, dreadfully, sometimes! F’r instance, here’s Tiny’s birthday next
week, and I’ve only twenty-five cents to buy her a birthday present
with, and she really needs a new doll; that old dud she carries about
isn’t fit to be seen, but what kind of a doll can you buy for twenty-
five cents? And then your birthday will be coming along, and then
papa’s and then Easter, and I want to give presents and send cards
to lots and lots of people, and how can I do it without any money?”
Mrs. Leslie could not help laughing.
“O Johnny, Johnny!” she said, “you’re as bad as the old woman
who called her lazy maids on Monday morning: ‘Come girls! Get up!
It’s washing day, and to-morrow’s ironing day, and Wednesday’s
baking day—here’s half the week gone, and you not out of bed yet!’
Dear little boy, we can’t have more than one day at a time, and here
you are borrowing trouble for almost a whole year!”
“Well, anyhow, mamma,” said Johnny, laughing in spite of himself,
and looking a little foolish, “Tiny’s birthday is, most here, and if I
might buy a quarter’s worth of peanuts, and sell them, and then
invest the money again, I do believe I’d have a dollar before it was
time to buy her present.”
“And I wonder,” said his mother, “how many of your lessons you
would learn, and on how many errands you would go for me, and
how many steps you would save for papa, when he comes home
tired, and how much carpentering you would do for Tiny and her
little friends? No, darling, if you can’t quite see what I mean, you
must just trust me. You can help a great many people, in a great
many ways, without money, and it is all beautiful practice for you,
against the time when you can help them with money too; but just
now, your main business is to see that papa and I are not
disappointed in the man that, with the dear Father’s help, we are
trying to help you to grow into. Keep your heart and your eyes open,
and you’ll see plenty of chances without the peanut-stand.”
Johnny looked, and felt, a good deal disappointed, but he was a
boy of his word, so he said resolutely,—
“I promised to trust you, mamma, and I will, for although you
never were a boy, papa was, and I sometimes think he’s a kind of
one yet; but you see I can’t help feeling pretty badly about it.
Perhaps it’s partly from sitting still so long—my legs are all cramped
up. Come out and race me just twice ’round the house,” he added,
coaxingly. “I should think your legs would be as stiff as pokers,
sitting sewing here the way you do, for half a day at a time!”
“They do feel a little stiff,” said Mrs. Leslie, springing up, and
dropping her sewing into the never-empty basket, “but for all that, I
think I can beat you yet, Mr. Johnny.”
She took off her apron and tucked up her skirt a little, and Johnny
made a line on the gravel-walk with a stick.
“Now, mamma, are you ready? One, two, three, off!” and away
they skimmed, down the walk, across the grassplot; into the walk
again, over the line, around once more, and then—
“There!” said Mrs. Leslie, triumphantly, “you’re beaten again,
Johnny Leslie!”
“I don’t care,” said Johnny, panting, and very red in the face,
“you’re only a foot ahead this time, mamma, and at that rate, I’ll be
two feet ahead, next time.”
The dinner-bell rang while Mrs. Leslie was smoothing her tumbled
hair and straightening her dress.
“I have an errand that will take me almost to the park this
afternoon, Johnny,” she said, at dinner, “Tiny is going with me, and if
you’d like to go, I will call for you at three, and ask to have you
excused from the writing hour, and then we can have a whole hour
in the park before we need come home to supper. Shall I?”
This was an extremely pleasing arrangement, and when the time
arrived, a happy party took seats in the horse car, for the park was
more than two miles from Mr. Leslie’s house, and the last part of the
way was decidedly an “up-grade.”
“Oh mamma!” exclaimed Tiny, “how will these two poor horses
pull such a car full of people up that steep hill? It’s too much for
them! Suppose we get out and walk?”
Tiny was always on the watch about the comfort of horses and
dogs and cats.
Just then the car stopped, and a third horse, that had been
standing patiently under a tree near the sidewalk, was fastened to
the pole in front of the other two, and, with his help, the car went
easily up the slope.
“That’s nice,” said Tiny, looking greatly
relieved, “I didn’t remember that they kept
an extra horse here, mamma; how good it
must make him feel, when the poor tired
horses stop and say, ‘That hill’s a great deal
too steep for us to drag this great heavy car
up it’; and then he says, ‘Hold on, I’m
coming. You can do it easily, with me to
help you!’”
“But, then,” said Mrs. Leslie, “just think
how much of his time he spends standing under the tree, doing
nothing but wait.”
“Why, mamma,” put in Johnny, “you know he knows the car will
be along presently, and while he’s waiting he’s resting from the last
trip, and getting up his muscle for the next one, so it isn’t exactly
doing nothing, even when he’s standing still.”
“And you don’t imagine that it makes him feel sorry that he hasn’t
a special car of his own to pull, but must just help other horses pull
theirs?” pursued Mrs. Leslie.
“I should think he’d be pretty foolish if he felt that way,” said
Johnny, confidently; “he’s doing something just as good, in fact, I
think perhaps it’s better, for he must make the two regular horses
feel good every time they come ’round there. Oh mamma, you’re
laughing! You’ve made me catch myself the worst ki—I mean
dreadfully! I see just what you mean; you might as well have said it;
you think that till I am old enough to have a car of my own, I ought
to be an extra horse!”
“But how could Johnny be a horse, mamma?” asked Tiny, deeply
puzzled.
They were out of the car by this time, and Tiny amiably joined in
the laugh which greeted this question.
“I’ll explain how he could when we sit down by the lake, darling,”
said her mother, “You and Johnny walk on slowly, now, while I stop
here for a few minutes and leave my work—the parcel, Johnny,
please!”
For Johnny was marching off with the parcel under one arm, and
Tiny under the other.
When they were comfortably seated on the shady green bank by
the lake, Mrs. Leslie explained to Tiny that she did not really expect
Johnny to turn into a horse, but that everybody who is looking out
for chances to help other people over their hard places will be sure
to find plenty to do.
“The world has a great many tired people in it,” said Mrs. Leslie,
“and a great many sick and sorrowful and discouraged and
disappointed people, and what a beautiful thought it is that the very
smallest and weakest of us may give help, and comfort, and
encouragement, every day of our lives, if we only will.”
“You do, mamma,” said Johnny, softly, stealing his hand into his
mother’s as he spoke, “and so does papa, but I’m afraid I’ve been
too busy with my own fun and things to try to help the poor tired
ones pull, but I truly mean to turn over a new leaf. I shall put it in
my prayers,” he added, reverently, and—“when, do you think, is a
good time for me to think, mamma? The time never seems to
come.”
“While you are dressing in the morning and undressing at night
would be very good times,” said his mother, “just before you say
your prayers, you know. You can think over in the morning what you
need most for that day, and at night what you have done and left
undone. I know your dressing and undressing don’t take long,” she
added, smiling, “but one can do a good deal of thinking in a few
minutes, if one gives the whole of one’s mind to it.”
The red sun, peeping under the tree beneath which they were
sitting, reminded Mrs. Leslie to look at her watch. It was high time
to start for home, and Tiny and Johnny, as the car went down the
steep hill, looked out with much affectionate interest for the “extra
horse,” and softly called good bye to him, as he stood quietly under
the tree, panting a little from his last pull, and patiently waiting for
the next.
I wonder how many of the dear little men and women who will
read this are training for their own life race by watching for chances
to help the hard-pressed runners who have started. Here is a motto
for all of you; the motto which a noble and earnest man has already
given to many people—“Look up, not down; look out, and not in;
look forward, not back; and lend a helping hand.”
And if you want his authority for this beautiful motto, it is easily
found, for you will all know where to look for these words,—
“Bear ye one another’s burdens, and so fulfil the law of Christ.”
CHAPTER VIII.
“LONG PATIENCE.”
iny and Johnny were planting their gardens,

and Jim Brady was helping them. Johnny
had happened to mention to Jim that he
liked a garden very well, after the things
were up, but that he did hate digging; and
Jim, after thinking hard for a minute, had
said,——
“See here! If you’ll teach me some of the
things you’re learning at school, of evenings, after my day’s work is
done, I’ll dig your garden for you, and do it better than you can, for
I’m a good sight stronger than you are, and I’ll help you keep it
clean all summer, too. Is it a bargain?”
Johnny hesitated. He did not like Jim’s tone. It was quite true that
Jim was the stronger of the two, but Johnny thought it showed bad
taste to mention it in that defiant sort of manner. And he did not see
any particular fun in teaching Jim, especially on summer evenings.
But it would be a great thing to have such good help with his garden
as he knew Jim would give, so he swallowed his pride, and said, as
graciously as he could,—
“All right. You come up after tea this evening, and we’ll begin. We
have tea at six, and I’ll hurry through mine, and then, when it’s too
dark to work any more, we can come into the playroom and have
the lesson.”
You will remember that it was this Jim Brady who had given
Johnny his first, and—there is reason to believe—his last cigar, and
so led him, though quite unintentionally, into his first act of deceit to
his mother. And the remembrance of this act was a very sorrowful
one, for although Johnny, as you know, had both confessed and
repented, and had been freely forgiven, the shameful act remained,
never to be undone. Do you ever think of that, when you are
tempted to do some mean, wicked thing?
Mrs. Leslie had called on Jim, at his bootblacking stand, soon after
this occurrence, and had a long talk with him, and the next time the
boys met, Jim had said, severely,—
“If I had an Angel for a mother, Johnny Leslie, I’d be shot before
I’d behave anyhow but on the square to her, and now I’ll put you on
your honor—if you find you’re learning anything she wouldn’t like,
from me, you’ve only to let me know, and I’ll cut you dead!”
This was a rather mixed statement, but Johnny understood it, and
felt himself blushing. It seemed to him that Jim had somehow got
things backward, but his recent downfall had humbled him, in more
ways than one, so instead of replying, as he was greatly tempted to,
that if anybody did any cutting, he would be the person to do it, he
merely said, rather shortly,—
“Very well, I guess I know a little more about my mother than you
do, so you attend to your mother-minding, and I’ll attend to mine!”
“Glad to hear it,” said Jim, easily, “but my mother’s what the
dictionary-talkers call a traydition; I never saw her, so I’d find it a
little impossible to mind her, don’t you see? But I’ll tell you one thing
—if your mother ever cares enough about me to give me a little
extra minding to do for her, I’ll see what I’m equal to in that line,
perhaps!”
Johnny had reported this speech to Mrs. Leslie, and she had
begun to work on the suggestion. Jim had already set his mark to a
promise not to smoke until he was twenty-one, and, although he did
not know it, Mrs. Leslie was trying to find him a situation where he
would have a certain, if small, salary, and be less exposed to
temptation than he now was. She was very glad when she heard of
the bargain which Johnny had made, and she presented the new
scholar with a slate and spelling book, at once. She also gave the
schoolmaster a little advice.
“You must remember, Johnny,” she said,
“that Jim has had no chance to learn
anything, compared with your chances, and
you mustn’t look superior, whatever you do.
Whenever you feel very grand, just imagine
how it would be if papa should write to you
in Greek, and talk to you in French and
Latin, and then call you a little stupid
because you could not understand him.”
Tiny looked rather mournful when she
heard of the new arrangement, but she brightened up, presently.
“Is he a very big boy indeed, Johnny?” she asked.
“Why, no,” said Johnny, considering, “at least, he’s not much
bigger than I am, Tiny. He’s only about half a head taller, but he’s a
good deal thicker.”
“What did you say you’d teach him?” pursued Tiny.
“Oh, all the things I’m learning at school, I s’pose!” replied Johnny,
“we didn’t settle about that, exactly, for I don’t know yet how much
he knows—he can’t write, but maybe he can read a little—I hope so,
for it must be awfully stupid work to teach people their letters. But
why do you want to know, Tiny?”
“I have a reason,” said Tiny, nodding her head wisely. “You
needn’t think you know all of everything, Johnny Leslie!”
“I never said I did!” retorted Johnny, warmly; then he looked at
Tiny, and began to laugh, she was so little, and was trying so hard to
look wise and elderly.
“You may laugh if you like,” she said, serenely, “I don’t mind. But
if you don’t know what you are going to teach him, perhaps you
know what you’re not. Are you going to teach him to sing?”
Johnny accepted Tiny’s gracious permission, and laughed a good
deal, but at last he answered,—
“No, Tiny, I’m not going to teach him to sing. I am quite sure
about that. Mamma says I can sing straight ahead first rate, but she
never knew me to turn a tune yet. I wish I could sing the way you
do,” he added, regretfully, “I’m so full of sing sometimes that I don’t
know what to do, but I can’t make it come out.”
They were sitting on the back porch, pasting their scrap-books,
and Mrs. Leslie was sewing at the window.
“Never mind, Johnny,” she said, consolingly, “you’ll not ‘die with all
your music in you’ while you do so much shouting.”
“Very well, then,” said Tiny, with a look of great satisfaction,
“when Jim comes, I shall tell him that if he will dig my garden for
me, I will teach him to sing.”
Mrs. Leslie expected to hear Johnny first laugh, and then try to
dissuade Tiny from carrying out her plan, but to her surprise, he did
neither. He said,—
“I shouldn’t wonder if he’d do it, Tiny; he’s all the time whistling,
and he whistles just like a blackbird, so very likely he’ll be glad to
learn to sing, too.”
When Jim came that evening, Tiny and Johnny were both in the
garden, and as Tiny had not yet met Jim, Johnny introduced them
thus,—
“Tiny, this is Jim. Jim, this is my sister Tiny, and she wants to be in
our bargain, too. Go ahead, Tiny.”
And so encouraged, Tiny went ahead.
“I have a garden, too,” she said, “but Johnny knows more of
everything than I do, except singing, and I thought perhaps you’d
like to learn to sing, and if you would, I’ll teach you that, and then, if
you think it is worth it, will you just do the hard digging for me? I
can do the rest myself, watching you and Johnny.”
A very gentle look came over Jim’s bold face, as he answered,—
“If you’ll teach me how to sing, Miss Tiny, it will be worth as much
to me as all Johnny can teach me of other things, and I’ll be proud
and happy to take charge of your garden.”
“Oh, thank you very much!” said Tiny, warmly. “What a nice, kind
boy you are! Do you mind if I watch you while you dig?”
“Not a bit!” said Jim, cheerfully, “I’m not bashful. But you’d better
sit down.”
“Wait a minute, and I’ll bring you your camp-chair, Tiny,” said
Johnny, and he raced to the porch for Tiny’s small chair, while Jim
pulled off the coat which he had put on as a mark of respect to Mrs.
Leslie, whom he hoped to see before the evening was over, and
went valiantly to work with the spade.
“What nice big spadefuls you make!” Tiny said, after watching him
a while. “When I dig, it ’most all slides off while I am picking up the
spade.”
“That’s because you are not quite so strong as I am,” said Jim,
smiling, and turning over an extra large spadeful by way of proving
his statement.
The two little gardens were thoroughly dug by the time that it was
too dark to work any more, and Johnny had hoed and raked Tiny’s
smooth, while Jim was digging his. Then they went into the
playroom, and Mrs. Leslie brought them a lamp to light up the
lesson.
“We will have a little singing first,” she said, opening the organ.
“Tiny and I will sing the evening hymn, and you must listen, Jim,
and try to catch the tune.”
Jim listened, and by the time they reached the Doxology, he had
joined them, and went through the tune without a mistake, seeming
even to know the words. His voice was a very sweet tenor, and Tiny
exclaimed delightedly,—
“It will be just as easy as anything to
teach him to sing, mamma!”
“I’d have come in sooner,” said Jim,
looking very much pleased, “but that last
verse was the only one I knew. I went to
Sunday-school a few times when I was a
little boy, and that verse came back to me
as soon as you began to sing it.”
Then Johnny and his pupil sat down by
the table, and Mrs. Leslie took Tiny’s hand
and went to the parlor, thinking that the two boys would manage
their undertaking better without an audience.
Johnny felt very much embarrassed, but he plunged in boldly, as
the best way of overcoming his feelings.
“I’ll do you the way they did me, the first day I went to school,” he
began, and taking his First Reader, he opened it, and handed it to
Jim, saying,—
“Just read a little, will you?”
Jim burst out laughing.
“It’s heathen Greek to me,” he said. “I don’t know more than half
the letters. Why, if I’d known how to read, I could have picked up
the rest somehow, and that’s why I asked you to teach me.”
Johnny was about to whistle, but he suddenly recollected his
mother’s warning.
“All right,” he said, composedly; “we’ll begin with the letters, and
I’ll teach you the way mamma teaches Tiny—it’s easier than the way
they do in school. Wait a minute, and I’ll borrow her card, the letters
are so much larger than they are in the spelling-book.”
He came back with a large card, covered with letters in bright
colors, and pointing to A, asked,
“Now, what does that look like to you?”
“It looks something like the tents those soldiers put up when they
camped near here,” said Jim, after looking at it for a moment.
“Very well; that’s A. Now, when you say ‘A tent,’ there you have it,
all right.”
“That’s easy enough to remember,” said Jim, “I thought it would
be harder.”
“I’ll tell you what this second fellow looks like, to me,” said Johnny,
delighted with Jim’s quickness, “it always makes me think of a
bumble-bee, and its name’s B.”
“That’s queer,” answered Jim, “it does
look like a big, fat bee, sure enough. I
guess I can remember that, too.”
It was not easy to find likenesses like
these for all the letters, but when Johnny
could not think of anything in the way of a
likeness, he told Jim of something strange
or funny that the letter “stood for,” and felt
quite sure, when the alphabet had been
“gone through,” that every letter was firmly
impressed upon Jim’s memory.
“Do you want to begin to learn to write now, or wait till you’ve
learned to read?” inquired Johnny, when the reading-lesson was
finished.
“I don’t know,” said Jim, “what’s the first thing you do when you
learn to write, anyhow?”
“You make ‘strokes’ first, like that—” and Johnny made a few
rapidly on the slate—“to sort of get your hand in, and then, when
you can make them pretty well, you go on to ‘pot-hooks and
trammels’—like these”—and he illustrated on the slate again—“and
when you can make them pretty well, then you begin to make
letters.”
“Well, then, I might as well begin right off,” said Jim, “I don’t have
to know how to read before I can make ‘strokes,’ that’s plain, and if
it takes so long just to get your hand in, the sooner I start, the
better!”
“Yes, I think so too,” said Johnny, encouragingly, “for of course,
you needn’t know how to read, to make ‘strokes’ or ‘pot-hooks and
trammels’ either, and you see you’ll be all ready, this way, to make
the letters, by the time you can read printing—maybe before. Here,
I’ll rule your slate, but I’ll ask mamma to set you the copy. I can’t
make as good strokes—or anything else for that matter—as she can,
and papa says a copy, any kind of a copy, ought to be perfect.”
Mrs. Leslie willingly set the copy, and guided Jim’s hand over the
first row. Nothing in her look or manner suggested to Jim that her
soft white fingers felt any objection to taking hold of his grimy ones,
but from that time he always asked Johnny for soap and water,
when the gardening was done, and came to his lessons with hands
as clean as vigorous scrubbing could make them.
When he had covered both sides of his new slate with “strokes,”
which Johnny assured him were quite as good as the first ones he
had made, they both decided that the lesson had been long enough
for that time, and parted with cordial good-nights.
“I didn’t know it was so easy to teach people, mamma!” said
Johnny, exultingly, as soon as his pupil was out of hearing, “why, it’s
no trouble at all!”
Mrs. Leslie smiled.
“Jim seems to be a bright boy,” she said, “but you must remember
that his mind is like your garden; things must be planted in it, and
you must wait a while for them to come up. I don’t wish to
discourage you, dear, but learning is a new business to him, as
teaching is to you, and I think this would be a good text for both of
you to start with—‘Let not him that girdeth on his harness boast
himself as he that putteth it off.’”
CHAPTER IX.
A CONTRACT.
three days’ rain which set in the morning

after Johnny’s first appearance as a
schoolmaster, put a stop to gardening, and
Jim decided for himself that he was not
entitled to any more lessons until he had
done some more work.
This had not been Tiny’s and Johnny’s idea
of the contract at all; they expected Jim to
help them whenever they needed help, and intended to keep on
regularly with their teaching, unless some very special engagement
should prevent them. But, as they remembered when they came to
talk it over, they had not made this plain to Jim, and they decided to
draw up a contract, and have it ready for his signature, or rather his
“mark,” if, as Johnny said rather mournfully, “it should ever clear up
again.” They lamented very much not having planted anything
before the rain.
“It would be soaking and swelling all the time,” mourned Johnny,
“and come bouncing up the minute the sun comes out!”
They tried shooting some radish seed at the beds with Johnny’s
pea-shooter, from an upstairs window, and had the pleasure of
seeing a flock of hungry sparrows make a breakfast of the seed
almost before it had touched the ground. Johnny was indignant, but
Tiny said tranquilly,—
“I’m glad I saw that. It was in last Sunday’s lesson, you know,
Johnny,—about the fowls of the air devouring it up. When things
don’t come up in my head, now, I shall know it was because I didn’t
plant them deep enough.”
It was after it had rained for two days and part of another, that
they drew up the contract, and thus it ran,—
“We are going to teach James Brady all we know,

that he wants to learn, and he is to come every
evening, unless we ask him not to, which we shall not
do except for something very particular, like a birthday
party, or having folks here to tea. And he is going to
help us work in our gardens, when we want help, but
he is to come all the same in the evening, whether he
has helped that day or not.
“Signed,
“Clementine and John Leslie
“James Brady.”
X HIS MARK
They admired this production so much, that they made

arrangements for framing it, when Jim should have added, “his
mark.” The arrangements consisted chiefly of an old slate-frame,
which Tiny painted bright red, using up her entire cake of vermillion
to do it, and Johnny was obliged to copy the contract in very large
letters, to make it fill the frame.
A day of brilliant sunshine followed the three days’ rain. Johnny
passed Jim’s stand on his way from school, reproached Jim for his
absence, told him of the contract, and secured his promise to come
that evening at a quarter past six, sharp. Tiny carefully practised a
little song for which she could herself play the accompaniment, and
both the children had their stock of seeds in readiness, before tea.
When Jim appeared, punctually at the appointed time, Mrs. Leslie
came out on the porch, and wished him good evening, and she
noticed with much pleasure that he had on a clean shirt, and that a
fresh patch covered the knee of his trousers, where a gaping rent
had been, four days ago. His face and hands shone with scrubbing,
and his hair with brushing, and he made the best bow at his
command, as he came up the steps.
“You’ll have to come too, mamma,” said Tiny, “for we haven’t quite
made up our minds where the things are to go, and we want you to
help us.”
“I’ll bring a camp-stool, and a board for your feet, mamma dear,”
chimed in Johnny, “and you can ‘sit on a cushion as grand as a
queen,’ and watch us work.”
“But I haven’t given papa his second cup of tea yet,” remonstrated
Mrs. Leslie, “nor eaten my piece of cake.”
“You can pour out the tea, and then ask papa to please excuse
you, and you can bring your cake with you,” said Johnny, coaxingly,
and to this Mrs. Leslie consented, although she said something
about tyrants. She came out, presently, with two pieces of cake on a
plate, and insisted upon Jim’s eating one of them, which he did
without the slightest reluctance, and then went vigorously to work.
You might have thought a large farm was being planted, if you had
heard the earnest discussion, and the number and variety of seeds
named, and dusk overtook them before they were half done. It was
decided that Tiny’s lesson should be given first, as her bedtime came
before Johnny’s did. The little song was quite new to Jim, and he
could not join in it as readily as he had joined in the hymn, but Tiny
went patiently over it, again and again, until he caught the air, and
knew the words of one verse, and she did not stop until they were
singing together in perfect harmony.
Then she gave him up to Johnny, and considerately left the room.
Johnny brought out the card with a flourish, saying confidently,—
“We’ll just run over the letters again, to make sure, and then we’ll
go on to the a-b-abs. Oh, here’s the contract—you just put your
mark to it there, where we’ve left a place, and then we’ll frame it
and give it to you.”
Jim listened thoughtfully, while Johnny read him the contract, but
he made no motion toward affixing his mark to it.
“It don’t seem to me to be fair,” he said, “you’ll not need much
work done in those little gardens, and here you’ve promised to teach
me nearly every evening; I think I ought only to have a lesson when
I’ve done some work.”
“Oh fiddlesticks!” said Johnny, impatiently, “you’ve worked like
everything already, and besides, we like to teach you; papa says it’s
the very best way to learn things, teaching them to somebody, so
you see it’s just as good for us as it is for you. Come, put your mark
there, where we left the hole for it,” and Johnny dipped the pen in
the inkstand, and handed it to his pupil, who reluctantly made his
mark in the “hole.”
“I’ll frame it to-morrow,” said Johnny, “Now for the letters. What’s
that?” and he pointed to V.
Jim pondered a moment, then,—
“That’s A,” he said, confidently.
Johnny controlled himself by a violent effort, pointed out the
difference between A and V, and then “skipped” Jim through the rest
of the alphabet. To his utter consternation, Jim only remembered
about half the letters, and of some of these he was not perfectly
certain.
“I didn’t think I was such a stupid,” said poor Jim, humbly, “but I
suppose that’s because I never tried to learn anything before. I
thought I knew half the letters before I began, but the boys must
have fooled me—I’m certain somebody told me that was K,” and he
pointed to R.
This made Johnny laugh, and Jim’s humility gave him such a
comfortable feeling of superiority, that he took courage, and went
through the alphabet once more, with tolerable patience. But Jim
was too keen-sighted not to notice the effort which Johnny was
making, and when the lesson was at last over, he said,—
“It’s going to be more of a job than you thought it would, Johnny;
I can see that, and if you want to be off your bargain, I’ve nothing
to say.”
But he looked so dull and disappointed, that Johnny’s conscience
reproached him with selfishness, and he said cheerfully,—
“Oh, you mustn’t give up the ship so soon, Jim. I’ll stick to it as
long as you will, and it will get easier after you’ve once learned the
letters. You’d better take your spelling-book home with you to-night,
and then to-morrow you can try to pick out the letters whenever you
have a little time, you know.”
“I will do that,” said Jim, brightening, “and I’ll not forget this on
you, Johnny—you’ll see if I do!”
Johnny went into the parlor, when Jim was gone, and dropped his
head on his mother’s shoulder.
“O mamma!” he said, dolefully, “he’d forgotten nearly every single
letter, and I could see he hardly believed me, when I told him that R
wasn’t K!”
Mrs. Leslie gently pulled Johnny down on her lap.
“You must go out bright and early to-morrow morning, and see if
your seeds are up,” she said.
Johnny looked at her in amazement.
“Why, mamma!” he exclaimed, “they’re only just planted! It will be
several days before they show the least little nose above ground.”
“Oh!” said Mrs. Leslie, but she said nothing more, only looking into
Johnny’s eyes with a little smile in hers.
He suddenly clapped his hands, exclaiming,—
“I see what you mean, mamma! I’m sowing seeds in Jim’s head,
and expecting to see them come up before they’re fairly planted! But
indeed, it’s harder work than digging.”
“‘Fair exchange is no robbery,’” said Mrs. Leslie, laughing at
Johnny’s mournful face. And then she said, quite seriously,—
“I will give you another text, dear; one that I thought of when I
was watching you plant your seeds this evening. ‘The husbandman
waiteth for the precious fruit of the earth, and hath long patience for
it, until he receive the early and latter rain.’ You see, the patience is
needed not only before the seeds come up, but while the plants are
blossoming, and while the fruit is forming, and while it is ripening. It
is not being patient just for a day, or a week, or a month, but for the
whole season, for it says ‘the early and latter rain.’ Now a great
many of us can have a little—a short patience, but it takes much
more grace to have the long patience, and this is what my little boy
must strive for.”
“I don’t think I’m naturally patient, mamma,” said
Johnny, with a sigh.
“No, I don’t think you are,” replied his mother, “but
Tiny is, and her patience will be a great help to you, if
you will only let it, just as your courage and energy are
a help to her, for she is naturally timid, and a little
inclined to be faint-hearted. You have a chance now to
win a great victory, and, at the same time, you are
running the risk of a great defeat; but you must not try
to have patience for the whole thing at once—ask
every day for just that day’s patience. You know when it is that we
don’t receive; it is when we ‘ask amiss.’ All our fighting for our Great
Captain will be in vain, unless we are ‘strengthened with all might,
according to His glorious power, unto all patience and long-suffering,
with joyfulness.’ We will see, next Sunday, how many times we can
find this word ‘patience’ in the Gospels and Epistles; you will be
surprised, I think, to find how often it is used.”
“It will be a help to remember, mamma,” said Johnny, with a more
hopeful look, “working in the garden, first. And I shall say ‘long
patience’ to myself ever so many times, before we begin our
lessons.”
So instead of going to bed with the discouraged feeling which the
lesson had left, Johnny went with a vigorous determination not to be
beaten, and he added to his evening prayer a petition for patience.
“If it hadn’t been for that contract, I wouldn’t have come a step
to-night,” said Jim, as they finished planting the gardens, the next
evening, “but I thought I would try one more shot, and then, if it’s
like last night, you must just let me off, and burn the contract up.”
“Indeed I shall not!” said Johnny, stoutly, “there it is, all framed
and glazed, and here I am, and there you are, and you’ll not get off
till you know how to read, and then you’ll not wish to!”
We will not follow Johnny through all the
discouragements and encouragements which
attended his career as a teacher; but you will be
glad to hear that, with that help which is always
near, he conquered, and that by the time he and Jim
were husking the corn which the little gardens had
yielded, Jim could read as fluently as his teacher
could, and was beginning to write a legible, if
somewhat uncertain hand. He had shown a real
talent for music, and, having learned all that Tiny
could teach him, was joyfully and gratefully taking
lessons from Mrs. Leslie.
“And just suppose my patience had turned out to be only the short
kind, Tiny!” said Johnny, as Tiny and he, with heads close together,
proudly popped the corn from their own farms.
CHAPTER X.
NEIGHBORS.
he desk next to Johnny’s had been vacant for

a long time, and he did not like this much,
for he was a sociable boy, and although of
course, no great amount of conversation was
permitted during school hours, it is
something to be able to make faces to a
sympathetic desk-mate. There was not an
absolute rule against talking in the school
which Johnny attended. The teacher had
said, at the beginning of the term,—
“Now, boys, I don’t forbid you to speak to each other during
school hours, if you have anything really worth saying on your
minds, and will speak so that you will not disturb your neighbors,
but all long conversations can be saved till school is out, and I hope
you will be honorable enough not to talk foolishly, or to take
advantage of this permission. If I find it necessary, I shall resort to a
rule, so you have the matter in your own hands.”
It had not been found necessary, so far, although the school was
full, excepting that one vacant seat next to Johnny’s.
“It may be a coincidence, you know, Tiny,” said Johnny, one day,
when he had been lamenting his lonely lot to his sister, “but I don’t
know—I have a kind of a sort of an idea that it isn’t.”
“What is a coincidence, anyhow, Johnny?” inquired Tiny, who was
never above asking for information.
“It’s two things happening together, accidentally, that look as if
they had been done on purpose,” explained Johnny, with the little air

Download Full Handbook of educational data mining 1st Edition Cristobal Romero PDF All Chapters

Uploaded by

Copyright:

Available Formats

Download Full Handbook of educational data mining 1st Edition Cristobal Romero PDF All Chapters

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Download Full Handbook of educational data mining 1st Edition Cristobal Romero PDF All Chapters

Uploaded by

Copyright:

Available Formats

Download the full version of the ebook at

Handbook of educational data mining 1st

Explore and download more ebook at https://ebookultra.com

The Handbook of Data Mining 1st Edition Nong Ye

Handbook of Statistical Analysis and Data Mining

Data Mining and Data Warehousing 1st Edition S.K. Mourya

Exploratory Data Mining and Data Cleaning 1st Edition

Music Data Mining 1st Edition Tao Li

Data Mining for Genomics and Proteomics Analysis of Gene

Data Mining for Bioinformatics Applications 1st Edition He

Practical Data Mining 1st Edition Monte F. Hancock Jr

AIMS AND SCOPE

No claim to original U.S. Government works

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

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

To my wife, Inma, and my daughter, Marta

To my wife, Ekaterina, and my daughter, Aleksandra

To my wife, Adriana, and my daughter, Maria

Part Iâ•… Basic Techniques, Surveys and Tutorials

2. Visualization in Educational Environments..................................................................... 9

3. Basics of Statistical Analysis of Interactions Data from Web-Based Learning

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

5. Classifiers for Educational Data Mining..........................................................................57

6. Clustering Educational Data...............................................................................................75

7. Association Rule Mining in Learning Management Systems.....................................93

8. Sequential Pattern Analysis of Learning Logs: Methodology and

9. Process Mining from Educational Data..........................................................................123

10. Modeling Hierarchy and Dependence among Task Responses in

Part IIâ•… Case Studies

12. Educational Data Mining to Support Group Work in Software

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

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

15. Automatic Recognition of Learner Types in Exploratory Learning

16. Modeling Affect by Mining Students’ Interactions within Learning

17. Measuring Correlation of Strong Symmetric Association Rules in

18. Data Mining for Contextual Educational Recommendation and Evaluation

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

20. Log-Based Assessment of Motivation in Online Learning........................................287

21. Mining Student Discussions for Profiling Participation and Scaffolding

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

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

24. Capturing and Analyzing Student Behavior in a Virtual Learning

25. Anticipating Students’ Failure As Soon As Possible...................................................353

26. Using Decision Trees for Improving AEH Courses.....................................................365

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

28. Lessons from Project LISTEN’s Session Browser.........................................................389

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

31. Machine-Learning Assessment of Students’ Behavior within Interactive

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

33. Using Markov Decision Processes for Automatic Hint Generation.........................467

34. Data Mining Learning Objects.........................................................................................481

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

The Purpose of This Book

Main Avenues of Research in Educational Data Mining

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

Dr. Cristóbal Romero is an associate professor in the

Dr. Sebastián Ventura is an associate professor in the

Dr. Mykola Pechenizkiy is an assistant professor in the

Dr. Ryan S. J. d. Baker is an assistant professor of psychol-

Wil van der Aalst Tiffany Barnes