Learning Analytics

This paper presents a novel approach and a method of learning analytics to study student agency in higher education. Agency is a concept that holistically depicts important constituents of intentional, purposeful, and meaningful learning. Within workplace learning research, agency is seen at the core of expertise. However, in the higher education field, agency is an empirically less studied phenomenon with also lacking coherent conceptual base. Furthermore, tools for students and teachers need to be developed to support learners in their agency construction. We study student agency as a multidimensional phenomenon centring on student-experienced resources of their agency. We call the analytics process developed here student agency analytics, referring to the application of learning analytics methods for data on student agency collected using a validated instrument. The data are analysed with unsupervised and supervised methods. The whole analytics process will be automated using microservice architecture. We provide empirical characterisations of student-perceived agency resources by applying the analytics process in two university courses. Finally, we discuss the possibilities of using agency analytics in supporting students to recognise their resources for agentic learning and consider contributions of agency analytics to improve academic advising and teachers' pedagogical knowledge.

As online training in the aviation industry continues to expand, understanding how learners engage in open online courses may help inform course design. Currently little is known about student engagement in aviation Massive Open Online Courses (MOOCs) and more information is needed about those who utilize, but do not complete them. Datasets from two aviation MOOCs were cluster-analyzed to determine subpopulations based on activity (discussions, videos, quizzes). Differences were examined for days of activity and completion. Survey data revealed differences in demographics and learning goals. Three significantly different subgroups were found for each MOOC. Engagement patterns were similar between corresponding levels across MOOCs for the most and least engaged groups, but differences were noted in the middle groups: MOOC 1 had a broader interest in optional discussions and videos, MOOC 2 had a narrower interest in optional discussions. Notably, significant associations were found between subgroups and days of activity, total quiz scores, and completions. In both MOOCs, significant differences were found between clusters and days of activity, with more highly engaged groups active more days than lower engaged groups. For both MOOCs, significant differences were found between cluster membership and total quiz score and significant associations were found between cluster group and course completion. In both MOOCs, the lower engaged clusters (Low and Moderate Engagers) showed a statistically significantly higher than expected proportion of students not completing the course.

In the past few years, the data analysis in the academic field has gained interest, because higher education institutions generate large volumes of information through historic data of students, information systems and the tools used by the learning processes. The analysis of this information supports the decision making, which positively impacts the academic performance of students and teachers. In addition, if the results of the analysis are shared with the institution community, it is possible to individualize the needs of the students and professors. Thus, the professional improving process can be applied to each person. This analysis needs to be focused on an education context such as Learning Analytics (LA) and Education Data Mining (EDM), which address the Knowledge Discovery process (KDP). Because of this, we present in this paper the implementation of LA in an Open Data (OD) environment to analyze the information of a higher education institution. The analysis is focused on the academic performance of the students from different perspectives (social, economic, family, among others). In order to improve the results, the analysis process is done in real time through Web Analytics (WA) for each member of the institution according to its needs.

Wenn digitale Informationssysteme im schulischen Kontext genutzt werden, entstehen auch Sammlungen von Daten zum Lernverhalten. Einige Anwendungen sammeln und analysieren diese Daten gezielt, um damit Lernende zu unterstützen. Das Forschungsfeld dazu heißt “Learning Analytics” und ist erst rund 10 Jahre alt (Leitner et al., 2019). In diesem Beitrag möchten wir zwei Anwendungen aus dem Schulkontext vorstellen, die Datenanalysen einsetzen um Lernfortschritte von Schüler/innen zu beobachten und Rückmeldung in Echtzeit zu geben (s. Ebner, Leitner, Ebner 2020). Wir nennen zudem Chancen und Herausforderungen von Learning Analytics im Schulkontext.

Utilitarian concerns increasingly drive machine ethics, particularly the few ethical discussions in learning analytics. Determining the most good for the most people is a responsible way forward, but perhaps not the most comprehensive, especially when assessing potential outcomes. Innovations in learning analytics may occur more rapidly and with better outcomes if informed ethical discussions occur at every step of development.

Learning Analytics ist ein sehr junges Forschungsgebiet aus dem Bereich der Bildungsinformatik und versucht durch das Sammeln und Gegenüberstellen von Daten Unterstützung im Lernprozess zu geben. Dieser Beitrag gibt eine Einführung und stellt die wesentlichen Aspekte, kritischen Punkte, Ziele und Methoden vor. Weiters soll dabei ein Bezug zur berufsbildenden Schulpraxis geschaffen werden.

I am a lecturer with professional marketing experience, and this study was motivated by my dismay at university about what I have perceived as oversights in colleagues’ typical use of mass email (emailing a single message to a large group of subscribers) for content delivery, in comparison to the communication strategies of major industries (for example, retail and public administration). I examine non-academic research about mass email best-practice, as implemented in industry communications; then I perform a systematic review of academic research on mass email for content delivery. My review of 19 studies confirms academics are behind the industry when it comes to mass email evaluation and optimisation strategies. Finally, I outline tactics necessary for more effective mass emails at university, the most salient being systematically analysing a range of email metrics, and testing and optimising emails. Also, included in this study is my identification of current obstacles to more effective mass email at university, and these include the technological limitations of learning management systems (LMS), and the limited expertise of teaching staff.

Learning Analytics ist eine Methode, die helfen kann, durch Datenanalyse das Lernen von SchülerInnen gezielt zu verbessern. Es werden Daten von einzelnen SchülerInnen generiert, zusammengeführt und analysiert, um daraus resultierend ganz individuelle pädagogische Interventionen setzen zu können, damit die SchülerInnen bestmöglich gefördert werden können (Ebner/Neuhold/Schön 2013: 1). Der Beitrag stellt das Konzept eines aktuell laufenden Forschungsprojekts vor. Einerseits soll durch einen Abgleich von aus Learning Analytics und Lernstandserhebungen gewonnenen Daten untersucht werden, wie wirksam die SchülerInnen unterstützt durch digitale Medien lernen. Andererseits ergeben sich durch das automatische Sammeln und Aufbereiten von Daten im Hintergrund der Software Möglichkeiten für LehrerInnen, die auch Einfluss auf die Vorbereitung und die Durchführung von leistungsdifferenziertem Unterricht durch datengestütztes und gezieltes individuelles pädagogisches Handeln der Lehrkräfte in der Volksschule haben können. Learning Analytics is a method that can help to improve the learning of students trough data analysis. Data from individual pupils are generated, merged and analyzed, so that as a result individual educational interventions can be undertaken by teachers to support students' individual learning (Ebner/Neuhold/Schön 2013: 1). On the one hand, by comparing learning analytics data and assessment results in this research project, it shall be found out how effectively students learn with the help of digital media.On the other hand, the automatic collection and processing of data in the background of the software offers opportunities for teachers, which may also influence the preparation and the classroom performance of teachers in differnciated settings through data-supported and individual pedagogical acting of teachers in elementary school.

Relevante Qualifikationsmerkmale bereits in der Vorauswahl erkennen zu können - ohne aufwändige psychologische Tests durchführen zu müssen - ermöglicht eine effektivere Preselektion von Bewerbern. In diesem Beitrag stellen wir ein integratives Modell der Qualifikation vor, das im Wesentlichen auf drei Grundkonzepten zur Qualifikation basiert und Alltagstheorien entspricht: dem psychologischen Konzept der Expertise, dem Konzept der Schlüsselqualifikation (Mertens 1974a&b) sowie dem Konzept von der Addition wissenschaftsgestützter Fachausbildungskomponenten, das der Ausbildungswirklichkeit an vielen Hochschulen entspricht. Die Stärke des Modells liegt in dem hohen Abstraktionsniveau der Dimensionen, die eine Zuordnung von Variablen aus unterschiedlichen Kontexten zum Modell ermöglichen. In einer Re-Analyse von drei Studien zum Umweltdienstleistungsmarkt in der Schweiz testen und zeigen wir: Mit diesem einfachen integrativen Modell von Qualifikation lassen sich (i) Branchenunterschiede aufzeigen (Qualifikationsprofile) und (ii) Berufserfolg vorhersagen.

Abstract. We investigated the validity of applying topic modeling to unstructured student text data from online class discussion forums to predict students’ final grades. Using only student discussion data from introductory courses in biology and economics, both probabilistic latent semantic analysis (pLSA) and hierarchical latent Dirichlet allocation (hLDA) produced significantly better than chance predictions which improved with additional data collected over the duration of the course. Predictions were more accurate from hLDA than from pLSA, suggesting the feasibility and value of deriving conceptual hierarchies relevant to actual student data. Results indicate that topic modeling of studentgenerated text may provide a useful source of formative assessment to support learning and instruction.

Dieser Beitrag ermöglicht eine kurze Einführung in das Themenfeld Learning Analytics mit einem besonderen Blick auf den Schulunterricht. Heute erscheint es noch weit entfernt, bis derartige Anwendungen im deutschsprachigen Raum flächendeckend Fuß fassen können. Durch die voranschreitende Technologie werden jedoch solche Anwendungen und die Auseinandersetzung mit der Frage, inwieweit künstliche Intelligenz Aspekte der eigentlichen Lehre ergänzen und erset-zen kann, zunehmend zum Diskussionsgegenstand. Die vorliegende Publikation zielt darauf ab, Learning Analytics selbst und die damit verbundenen Herausforderungen zu definieren. Anschlie-ßend werden einige allgemeine Beispiele genannt, ehe auf zwei webbasierte Informationssysteme im Detail eingegangen wird-dem Einmaleins-Trainer und dem Programm zum Aufbau von Schreibkompetenz IDeRblog. Auf Basis der dort gewonnen Erkenntnisse und Erfahrungen werden drei wesentliche Anforderungen für Lehrerinnen und Lehrer abgleitet: statistische und digitale Kom-petenz sowie grundsätzliches Wissen im Bereich Datenschutz. Der Beitrag schließt mit der Frage, inwieweit diese zukünftig in die Lehrerbildung integriert werden können und müssen.

Περίληψη Ένα νέο πεδίο που διευκολύνει την εκπαιδευτική διαδικασία και βελτιώνει την απόδοση των εκπαιδευομένων, προσελκύει τα τελευταία χρόνια τους εκπαιδευτικούς αναλυτές και ερευνητές. Είναι τα learning analytics (LA) που αφορούν τη χρήση των μαθησιακών δεδομένων τα οποία συμβάλλουν στη βελτίωση των εκπαιδευτικών πρακτικών και την κατανόηση της μαθησιακής διαδικασίας. Σε αυτό το άρθρο παρουσιάζεται ο ρόλος που διαδραματίζουν στην τριτοβάθμια εκπαίδευση, βασισμένος σε διεθνή βιβλιογραφική ανασκόπηση επιστημόνων της εκπαίδευσης. Παρουσιάζονται επίσης 12 εμπειρικές μελέτες-ερευνητικά ερωτήματα, μεθοδολογία και ευρήματα-που έγιναν τον τελευταίο χρόνο στο πεδίο των learning analytics. Αρχικά, δίνεται περιγραφή του ορισμού του νέου πεδίου, κατόπιν γίνεται μία ειδική αναφορά στη χρήση του στα ανώτατα εκπαιδευτικά ιδρύματα. Στη συνέχεια, παρουσιάζεται η μεθοδολογία που ακολουθήθηκε, τα ερευνητικά ερωτήματα και τα ευρήματα των βιβλιογραφικών ερευνών. Κατόπιν δίνονται τα πλεονεκτήματα των learning analytics και οι περιορισμοί τους οποίους αντιμετώπισαν οι ερευνητές.To άρθρο ολοκληρώνεται, με τη διεξαγωγή και διατύπωση των συμπερασμάτων και τη παράθεση προτάσεων για μελλοντική έρευνα των learning analytics στην εκπαίδευση γενικότερα. Abstract A new field that facilitates the educational process and improves the performance of learners attracts educational analysts and researchers in recent years. Learning analytics (LA) is the usage of learning data that helps to improve educational practices and understanding the learning process. This paper reveals an international bibliographic review of empirical researches of education scientists on the usage of learning analytucs in higher education. I also present empirical studies-research questions, methodology and findings – made of the last year in the field of learning analytics. Firstly, presented a description of the definition of the new field and then is given a special reference to its usage in higher education institutions. The nest section is the methodology followed, the research questions and the findings of bibliographic research. Then examined benefits of learning analytics and limitations that researchers faced are given. Finally, this paper concludes by making and formulating the conclusions and proposing a future study of learning analytics in education in general. Λέξεις κλειδιά: learning analytics, τριτοβάθμια εκπαίδευση, μεγάλα δεδομένα A.Εισαγωγή Στοχεύοντας στην κατανόηση και στη βελτιστοποίηση της μάθησης και του μαθησιακού περιβάλλοντος όπου αυτή πραγματοποιείται, οι εκπαιδευτικοί αναλυτές αναπτύσσουν νέα εργαλεία για τους εκπαιδευόμενους και τους εκπαιδευτικούς. Προκειμένου να αναπτυχθούν εργαλεία για τη διευκόλυνση της εκμάθησης, ένα σημαντικό πρώτο βήμα είναι να κατανοηθούν οι στάσεις των εκπαιδευομένων και οι ανησυχίες τους (Roberts, 2016). Στη συνέχεια, να υιοθετηθούν πολιτικές που έχουν ως στόχο να εξασφαλίσουν την αποτελεσματικότητα και την καταλληλότητα (Tsai et al., 2017). Η πρόβλεψη της απόδοσης ενός μαθητή είναι ωφέλιμη για τη συμμετοχή του στο εκπαιδευτικό σύστημα. Οι υπάρχουσες εκπαιδευτικές μέθοδοι χρησιμοποιούν χαρακτηριστικά που συνδέονται κυρίως με την ακαδημαϊκή απόδοση, το

Résumé

Cette thèse a pour objectif d’étudier la faisabilité de différentes techniques d’amplification des données textuelles (ADT) basées sur le traitement de la langue naturelle (TLN) et l’apprentissage automatique afin de pallier l’insuffisance de données pour l’entraînement de modèles en apprentissage profond. Pour y arriver, nous montrerons une différence de performance entre les mêmes modèles entraînés avec et sans données amplifiées.

Sur le plan pratique, il est fréquent de se retrouver avec des quantités insuffisantes de données pour entraîner de gros modèles statistiques. Ce «mur des données massives» représente un défi à la fois pour les communautés linguistiques minoritaires sur la Toile, les organisations, les laboratoires et les entreprises qui rivalisent avec les géants du GAFAM.

Alors que le gros de l’effort de recherche en amplification textuelle se concentre sur des solutions d’apprentissage de bout-en-bout, le présent travail s’oriente plutôt vers l’emploi de techniques de prétraitement des données pratiques, robustes, capables de monter en charge et simples à mettre en oeuvre. Ces techniques sont inspirées par les techniques d’amplification utilisées avec succès en vision artificielle.

L’idée est de créer de nouvelles données à partir des données existantes. Par abus de langage on parle d’augmentation des données, mais il s’agit plutôt d’une amplification, puisque l’on crée de nouvelles données en préservant le sens qui demeure invariant. On parle aussi de données synthétiques, de données générées ou de données artificielles.

Bien que les techniques étudiées s’appliquent à tout genre de textes, cette thèse se concentre sur l’amplification de phrases. Nous verrons que le concept de «transformation sémantiquement invariante» est au coeur du processus d’amplification des données. 

Plusieurs techniques d’ADT ont été expérimentées. Certaines furent testées pour fins de comparaison comme l’injection de bruit ou l’emploi d’expressions régulières. D’autres ont été améliorées comme la substitution lexicale. D’autres techniques plus innovatrices font appel à des services en ligne robustes et capables de monter en charge comme la rétrotraduction et la génération de paraphrases par la transformation d’arbres syntaxiques.

Les techniques d’amplification textuelle étudiées ont permis d’accroître l’exactitude des résultats dans une fourchette 0.5 à 8.8%, sur une tâche normalisée de prédiction de la polarité de textes de critiques de film de la base de données IMDB. Différentes architectures de réseaux profonds de neurones ont été testées: le perceptron multicouche (PMC), le réseau de neurones convolutif 1D (RNC 1D), le réseau récurrent à longue mémoire court terme (LMCT) et le réseau récurrent LMCT bidirectionnel (biLMCT).

Les techniques de transformation de surface (bruit lexical, fautes d’orthographe, expressions régulières) s’avèrent les plus performantes compte tenu de leur rapidité et de leur faible besoin en calcul. La substitution lexicale avec des vecteur-mots AdaGram qui traitent la polysémie est meilleure que la substitution lexicale simple avec WordNet. La génération de paraphrases basée sur la rétrotraduction s’illustre particulièrement sur certains jeux de données et mériterait d’être approfondie.

La principale limite de ce travail est que l’expérimentation ne réalise qu’une seule tâche, soit la prédiction de la polarité de textes. Aussi l’objectif se limitait à montrer la faisabilité des différentes techniques d’ADT. Par exemple, les expériences se sont limitées à un facteur d’amplification de cinq (5), soient cinq nouveaux textes par texte original ce qui ne représente, dans bien des cas, qu’une infime partie des combinaisons possibles.

Malgré leur côté simple et pratique qui plaira aux praticiens, le principal inconvénient de certaines des techniques d’ADT explorées est la quantité de calculs requis. Notons aussi la dépendance envers les services de traduction et d’analyse syntaxique de fournisseurs privés. Nous proposons des alternatives pour mitiger cette dépendance.

En conclusion, nous faisons quelques propositions de futurs travaux, des pistes de valorisation et discutons de l’émergence d’un nouveau défi, le «mur des modèles préentrainés», qui combine à la fois la masse des données et la masse des calculs.

Long-tested pedagogical theories can provide the basis to analyze and interpret students’ interactions in distance learning environments. Through this lens, useful conclusions can be drawn from students’ data that could help to improve the teaching and learning process. In this paper, a number of forum posts were analyzed and the role of tutors and students was scrutinized through the use of network visualization. It was found that certain students have a central positive role in the discussion forum although tutors have the main burden of answering and keeping the interaction alive. These students can act as the more knowledgeable other and support their peers. Further research in order to gain insight into students’ interaction and promote collaboration, is proposed.

In this paper we describe some case studies of the use of virtual worlds in corporate training as well as Higher Education. In particular for Higher Education we describe how the Virtual World constructed using the platform Avaya Live Engage, is used as an immersive environment with pre-service teachers, who are undergoing a 1-year teacher training program, and how the data analytics collected in-world is being used to monitor and direct content development. We focus our studies on the initial hypothesis that 3D immersive environments are highly engaging and offer an experience that goes beyond the 'traditional' online education. We want to combine different analysis methods to be able to get empirical evidence showing the students' engagement with the 3D space in ways that can help us in the design of the learning experience accompanying the learners in their journey. In this paper we describe the research methods we use for the study, and give an overview of the information we can collect from the in-world analytics. We also propose how these analytics can be used for a predictive model with the intention of refocusing the virtual world experience to match learner needs.

Abstract: The growing number of on-line interactions between learners and on-line learning systems leaves a trail of data that can be analyzed at a number of levels of granularity and for several purposes. Learner engagement in the program of studies, with the course or with a specific learning activity is one variable thought to highly correlate with learner success. Engagement may be with the learning content, or with other learners and faculty in the socially constructed environment. This paper looks at three specific areas where learner engagement can be measured, and discussed the possibilities for using such information in advising learners and faculty on ways to strengthen course outcomes.
Keywords: Learning analytics, Learner engagement, student success, on-line learning

Workshop content:

The workshop will focus on the utility of Coh-Metrix in discourse theory and educational practice. We will begin with an introduction to the Coh-Metrix and Test Easability Assessor (TEA) tools, including a detailed description of the indices in the context of multilevel theoretical frameworks of discourse comprehension and learning. A review of published studies will help illustrate the relationship between discourse and learning. We will then walk participants through a series of practical exercises that will give hands-on experience with cleaning, preparing, and analyzing texts using Coh-Metrix, conducting statistical analyses with Coh-Metrix data, and data interpretation.

Intended learning outcomes:

Participants will become familiar with the structure of the Coh-Metrix and Test Easability Assessor (TEA) tools and their applications, with particular focus on the analysis of educational data. Participants will learn about the various indices available in each tool, and how to interpret their significance.

The analysis of data collected from the interaction of users with educational and information technology has attracted much attention as a promising approach for advancing our understanding of the learning process. This promise motivated the emergence of the new research field, learning analytics, and its closely related discipline, educational data mining. This paper first introduces the field of learning analytics and outlines the lessons learned from well-known case studies in the research literature. The paper then identifies the critical topics that require immediate research attention for learning analytics to make a sustainable impact on the research and practice of learning and teaching. The paper concludes by discussing a growing set of issues that if unaddressed, could impede the future maturation of the field. The paper stresses that learning analytics are about learning. As such, the computational aspects of learning analytics must be well integrated within the existing educational research.

In 2016, Facebook founder Mark Zuckerberg announced that he would be joining the National Science Foundation and Gates Foundations in spending hundreds of millions of dollars to support the development of technologies for “personalized learning.” These technologies promise to provide learners or users with interactive content that is tailored with precision to their preferences and previous behaviors—much like Facebook’s newsfeed does. And also like Facebook, they accomplish this by using “big data” and “analytics,” recordings of clicking, scrolling and typing of many thousands of users. However, analytics experts, sociologists of technology and most recently, former Facebook and Google executives understand these techniques as ultimately representing a sophisticated and highly effective form of Skinnerian operant conditioning—one that customizes schedules of reinforcement precisely to users’ or learners’ behavioral patterns. First explaining how this “new behaviorism” is instantiated in personalized learning design and technology, this paper then argues that such an approach is antithetical to the most basic priorities and purposes of education: Namely, to cultivate in students a sense of ownership in their own learning, and responsibility for their own behavior and its effects on others.

The Handbook of Learning Analytics is designed to meet the needs of a new and growing field. It aims to balance rigor, quality, open access and breadth of appeal and was devised to be an introduction to the current state of research. The Handbook is a snapshot of the field in 2017 and features a range of prominent authors from the learning analytics and educational data mining research communities. The chapters have been peer reviewed by committed members of these fields and are being published with the endorsement of both the Society for Learning Analytics Research and the International Society for Educational Data Mining. We hope you will find the Handbook of Learning Analytics a useful and informative resource.

New learning technologies require designers and faculty to take a fresh approach to the design of the learner experience. Adaptive learning, responsive and predicative learning systems are emerging with advances in learning analytics, the process of collecting, measuring, analysing and reporting data with the intention of optimising the student learning experience itself and/or the environment in which the experience of learning occurs. However, it is suggested here that no matter how sophisticated the learning analytics platforms, algorithms and user interfaces may become it is the fundamentals of the learning, grounded in meaningful pedagogical and andragogical theories of learning that will ensure that technology solutions will deliver significant and sustainable benefits.

In 2016, Facebook founder Mark Zuckerberg announced that he would be joining the National Science Foundation and Gates Foundations in spending hundreds of millions of dollars to support the development of technologies for “personalized learning.” These technologies promise to provide learners or users with interactive content that is tailored with precision to their preferences and previous behaviors—much like Facebook’s newsfeed does. And also like Facebook, they accomplish this by using “big data” and “analytics,” recordings of clicking, scrolling and typing of many thousands of users. However, analytics experts, sociologists of technology and most recently, former Facebook and Google executives understand these techniques as ultimately representing a sophisticated and highly effective form of Skinnerian operant conditioning—one that customizes schedules of reinforcement precisely to users’ or learners’ behavioral patterns. First explaining how this “new behaviorism” is instantiated in personalized learning design and technology, this paper then argues that such an approach is antithetical to the most basic priorities and purposes of education: Namely, to cultivate in students a sense of ownership in their own learning, and responsibility for their own behavior and its effects on others.

Discussion boards can provide a glimpse into the regular and substantive interaction required in online courses. Advances in technology and an increased interest in learning analytics now provides researchers with billions of data points about instructor and student interaction within a learning management system (LMS). This study used LMS data to explore the frequency of interaction between instructors and students in discussion boards in online courses at one institution. Overall, 415 courses were analyzed spanning two semesters. Results from the study found that the average number of posts by an instructor was 32.9. The average instructor interaction was 1.49 instructor posts per student. 23% of courses had no instructor posts. Student posts averaged 470 per course and the average posts per student was 19.9. Based on the discussion board activity, the most discussion interaction occurred during the first two weeks of the semester. Results also suggested that there is no relationship between student satisfaction and the number of total posts in a course. The paper concludes with implications for research and practice.

Digital communication has changed human life since the invention of the internet. The growth of E-mail, social websites and other interpersonal communication systems in turn have brought rapid development in especially the key technological area of data analytics. Using advanced forms of analytics helps the examination of data and better informs investigative sense-making and decision-making of all kinds. The legal process called Electronic discovery (E-discovery) is used for investigating various events in the digital communication world, for the purpose of producing/obtaining evidence (such as evidence in the form of emails used in the Enron fraud case). Investigating digital communications collected over a period of time, manually, is a strenuous process, time consuming, expensive and not very effective. More recently, within E-discovery there has been development of analytics known in the legal community as “Technology assisted review” (TAR). TAR is a technologydriven assistant ...

Big data in education has fostered emergent fields like educational data mining (Baker & Yacef, 2009) and learning analytics (Siemens & Long, 2011). Simulations and educational videogames are obvious candidates for the application of these analytic methods, affording big data situated in meaningful learning contexts (Gee, 2003; Steinkuehler et al., 2012). In design of these educational games, clickstream analytics for core design, alpha usertesting, and final-stage adaptive play design play a key role in optimizing learner experience. This paper maps learning analytics methods to these learning game development phases. Leveraging these powerful analytic tools of visualization, association mining, and predictive modeling throughout the design process is key to supporting players in a user-adaptive, engaging play experience optimized for learning.

This paper discusses the general thesis that massive open online courses (in short MOOC), open educational resources (in short OER) and learning analytics are an impactful trio for future education, especially if combined. The contribution bases upon our practical experience as service providers and researchers in the department "Educational Technology" at Graz University of Technology (TU Graz) in Austria. The team members provide support to lecturers, teachers and researchers in these addressed fields for several years now, for example as host of the MOOC platform iMooX.at, providing only OER since 2015. Within this contribution, we will show, against some doubtful or conflicting opinions and positions, that (a) MOOCs are opening-up education; (b) learning analytics give insights and support learning, not only online learning, if implemented in MOOCs; and (c) that OER has the potential for sustainable resources, innovations and even more impact, especially if implemented in MOOCs.

Проанализирован опыт Томского государственного университета по обеспечению качества электронного обучения на основе показателей качества учебных материалов и сопровождения учебного процесса. Возможности управления этими показателями рассмотрены на основе системы управления учебным процессом MOODLE с применением базовых инструментов и специальных сервисов для LMS MOODLE, разработанных вТГУ: общая статистика по учебным подразделениям, содержание и настройки электронных учебных курсов, активность студентов и преподавателей, промежуточные результаты обучения, обращения пользователей в службу поддержки.Авторыработыопределяютизмеряемые показатели, описывают методику обработки и анализа полученных данных для принятия управленческих решений.

A Analítica da Aprendizagem (do inglês Learning Analytics - LA) é um campo de estudos que tem o objetivo de entender e aprimorar os processos de ensino e de aprendizagem por meio de coleta, análise e ação sobre dados digitais que tenham alguma relação com o contexto educacional. Em cursos a distância, os ambientes virtuais de aprendizagem (AVA) podem ser usados como meios para realizar a coleta de dados que tenham relação com o desempenho educacional dos alunos, haja vista que o processo educacional ocorre no próprio ambiente. O Moodle é um importante AVA e uma das ferramentas mais utilizadas para obtenção de dados na LA; no entanto, não dispõe de recursos em sua instalação-padrão que permitam realizar análises mais avançadas sobre os dados do próprio ambiente. Uma ferramenta relevante para uma das técnicas que a LA utiliza, a Mineração de Dados Educacionais (MDE), é o WEKA, que realiza análises por meio de algoritmos diversos de aprendizado de máquina. Com base nisso, esta pesquisa busca responder o seguinte problema: como deve ser um plugin para o AVA Moodle que, tomando por base técnicas da LA, utilize algoritmos da ferramenta WEKA para realizar ações de monitoramento, análise, reflexão, predição e intervenção relacionadas aos estudantes? Um plugin para esse fim foi desenvolvido e fundamentado no modelo de referência para uso da LA proposto por Chatti et al. (2012), utilizando os conceitos da Arquitetura Orientada a Serviços, com a tecnologia de Web Services, para permitir a integração do WEKA com o Moodle. Foram disponibilizadas ferramentas como relatórios, painéis, gerenciamento de intervenções e alertas, objetivando prover de maneira simples e intuitiva, aos educadores e alunos, recursos para realizar monitoramento, análise, reflexão e predição do desempenho acadêmico de estudantes, além de intervenção na situação encontrada. As ferramentas foram avaliadas com base em dados de disciplinas do curso de Licenciatura em Matemática da Universidade Estadual do Ceará (UECE), em parceria com o Sistema Universidade Aberta do Brasil (UAB), no polo de Caucaia-CE, no sentido de verificar se as ferramentas atendem ao modelo de referência da LA, e de analisar a qualidade dos modelos de previsão de desempenho gerados. Como resultado, observou-se que o modelo de referência foi atendido e os modelos gerados possuem boa capacidade de discriminação desde a quinta semana de aula. Seria útil, entretanto, que houvesse melhor padronização no planejamento das disciplinas e melhor uso dos recursos do Moodle, para permitir previsões mais precisas.