|9:00 – 9:30
||Opening and welcome
||9:00 – 10:30
||Workshop “Educational data analysis using R” by Pedro J. Muñoz Merino
|9:30 – 11:00
||Keynote “Visual Learning Analytics in Computing Education” by Sharon HSIAO
|10:30 – 12:00
|11:00 – 12:15
||From Big Data to Learning Analytics, an industrial approach
|12:00 – 12:30
|12:15 – 12:45
|12:30 – 14:00
|12:45 – 14:00
||Workshop “Enhancing educational data quality in heterogeneous learning contexts using Pentaho Data Integration” by Alex Rayón
|14:00 – 15:30
|14:00 – 15:30
||Bilboats 1 hour boat trip
||Visit to the Guggenheim Museum
KEYNOTE “A Learning Analytics MOOC: What we’ve learned” by Ryan Baker
In 2013, I taught my first MOOC, Big Data and Education, on the Coursera platform. Since then, this MOOC and two successor MOOCs have been a useful tool for my lab’s learning analytics research, for intervention studies, and for flipping my educational data mining methods course at my university. In this talk I’ll discuss our findings, in areas as disparate as communities of practice, online discourse, emotion and negativity online, and retention in both MOOCs and traditional graduate courses.
KEYNOTE “Visual Learning Analytics in Computing Education” by Sharon HSIAO
Engaging students with social learning technologies has become an important trend in modern e-learning. One of the biggest challenges is to provide adaptation in the context of social learning, while at the same time allowing students to feel in control. I have begun studying and extending the benefits from adaptive navigation support to broader audience by incorporating social learning technologies. I designed and studied various visual approaches of open student modeling interfaces that provide students with easy-to-grasp and holistic views of their progress and knowledge. I call it Open Social Student Modeling visualizations. Specifically, I focused on the visual representations in the context of social learning, which can be considered as Visual Learning Analytics. Results revealed that combining personalized and social learning technologies through social visualizations enhanced the learning quality in increasing their motivation and engagement (Hsiao & Brusilovsky, 2011; 2012; Hsiao, et al., 2011; 2012; 2013). In the talk, I’d like to also present my recent work on semantic visual analytics in the same context of online programming language learning.
Workshop “Enhancing educational data quality in heterogeneous learning contexts using Pentaho Data Integration” by Alex Rayón Jerez
Heterogeneous data integration and quality normalization are sensitive issues to properly exploit learning data. In this hands-on tutorial, we will not only extract learning data from their databases, but also enhance data quality issues (granularities, dimensions, duplications, nulll values, etc.) through the use of Pentaho Data Integration. We will practice with the integration of learning data from technology-rich learning environments (LMS, Social Networks, wiki, etc.). It is required the use of a laptop with Pentaho Data Integration module already installed on it, but it is not required previous knowledge of Pentaho
Workshop “Educational data analysis using R” by Pedro J. Muñoz Merino
This hands-on tutorial is focused on the practical use of the R free software statistical tool in order to analyze educational data
sets. We will provide examples on the use of R for hypotheses testing, prediction and clustering in educational
scenarios. It is required the use of a laptop but it is not required previous knowledge of R.
SESSION: From Big Data to Learning Analytics, an industrial approach
- Norberto Ojinaga Goitia, Engineering and Technological Planning Director, Euskaltel, www.euskaltel.com
- Elisa Martín Garijo, Directora de Tecnologia e Innovación de IBM, www.ibm.es
- Liliana Arroyo Moliner, PhD. Eticas Research & Consulting, www.eticasconsulting.com
SESSION: Experiences I
- Jordi Cuadros, Assessing open-ended learning activities through analyzing action logs, ASISTEMBE – IQS Universitat Ramon Llull
- Abstract. Although open-ended activities are often thought as critical to promote critical thinking and problem-solving abilities, these are not easy to assess. When these open-ended activities are computer-based, relevant information can be retrieved from the action logs both to know whether the activity is effective and to grade each student’s work. In this contribution, we will present how we analyze action logs from a few different open-ended educational applications in chemistry (a virtual lab), in physics (a remote lab), in statistics (using R Commander) and in business (Excel-based activities) and what information can be obtained to facilitate these assessment efforts.
- Juan Cruz Benito, VeLA: A Visual eLearning Analytics tool, GRIAL research group. University of Salamanca
- Abstract. VeLA is a Visual eLearning Analytics model. This presentation is devoted to introduce the tool that supports this model. Specifically, we are going to underline the four visualizations that have been developed to represent 1) the full overview of an entity, both virtual campus or specific course, based on parallel coordinates; 2) the social interaction graph inside the course; 3) the usage time patterns; and 4) the summary of the main terms used in the coursed in a word-cloud map. The visual analytics applied to the eLearning establishes a analysis flow that combines both the automatic processes supported by the computers and the human experience to find knowledge patterns interacting with the data visualizations.
- Mr Kais Dai. EACEA Erasmus Mundus Doctoral ScholarShip, Applications of Social Data Mining to Learning Analytics, Information & Computing Lab, Atlantic Research Center for ICT. Universidad de Vigo (SPAIN)
- Abstract. The emergence of Massive Open Online Courses (MOOCs) as a new learning alternative for lifelong learners highlights the need of a more personalized learning experience especially regarding the offers diversity and abundance. Also, the dynamics of the job market inherits new interactions’ models among the different stockholders (recruitment companies, job seekers, etc.). In this context, analysing the job market needs on OSNs is a promising issue, so, the objective here is to take advantage of collaborative data sources (OSNs) in order to provide a personalized recommendation’s system for lifelong learners (MOOCs users) and enhance their career path according to an objective driven approach guided by the job market needs. In this sense, we are applying social mining techniques to merge and process a big amount of data to discover and characterize OSNs’ users, to analyze and detect the job market needs and MOOCs offers. The idea is to merge all these data sources and be able to recommend MOOCs and warn both learners and online courses providers to react properly.
- Mariluz Guenaga, DeustoTech Learning, working on the assessment of generic and specific competencies using Learning Analytics, DeustoTech Learning – Faculty of Engineering, University of Deusto (Spain)
- Abstract. DeustoTech Learning is a research group at the Engineering Faculty of the University of Deusto. We work on improving learning and teaching through technology, focused on STEM (Science, Technology, Engineering, Mathematics) areas. Learning Analytics is a transversal area of research, along with Remote Experimentation and Game-based learning. Many of our projects include Learning Analytics to support the assessment of generic and specific competencies. In this talk we will show the work we are doing in ongoing projects such as Auto Game, Kodetu, Visir Remote lab and SCALA.
SESSION: Experiences II
Pedro J. Muñoz Merino, Introducing Learning Analytics in High School Education for Adults, Universidad Carlos III de Madrid, CEPA Sierra Norte de Torrelaguna
- Abstract. Massive Open Online Courses (MOOCs) can enable flipped classroom methodologies, in which students can watch videos, solve exercises and perform other activities online before the face to face lessons. This methodology has been introduced in High School for adults in the CEPA Sierra Norte de Torrelaguna with the Open edX platform for a MOOC in maths. In this scenario, the use of learning analytics to monitor and better understand the students interactions with the MOOC can improve the learning process. We present ANALYSE, a learning analytics tool developed at Universidad Carlos III de Madrid for the Open edX platform, which provides 12 new visualizations for teachers and students. In addition, we explain how ANALYSE has been useful for self-reflection, providing feedback or detecting issues with educational resources in a MOOC of maths in High School education for adults with a flipped classroom methodology. This experience is within the framework of the mapatic and emadrid projects.
- Alejandra Martínez Monés, Aligning learning analytics and learning design to support teachers in orchestrating CSCL situations – GSIC/EMIC, Universidad de Valladolid
- Abstract. Aligning learning analytics and learning design to support teachers in orchestrating CSCL situations – GSIC/EMIC, Universidad de Valladolid
From the conceptualization to the evaluation of computer-supported collaborative learning (CSCL) scenarios, teachers address multiple tasks, sometimes being overwhelmed on account of the required time and associated burden. To support teachers in this endeavor, we propose to connect the pedagogical decisions made at design time with the analysis of the participants’ interactions. Thus, teachers would be provided with relevant and coarse-grained information that could help them manage their CSCL scenarios. This talk synthesizes the main contributions obtained from a 3-year design-based research process, and presents the findings obtained from the evaluation of the current proposal in two authentic CSCL scenarios. The participant teachers valued the proposal positively and stated that it was helpful for their orchestration of CSCL scenarios.
The talk is based on the work by M.J. Rodríguez Triana (2014), Linking scripting and monitoring support in blended CSCL scenarios, PhD thesis dissertation, UVa. Available at: http://uvadoc.uva.es/handle/10324/7049
- Jordi Sancho, Análisis de Sistemas de Aprendizaje Colaborativo a Gran Escala Basados en Wikis, Laboratorio de Medios Interactivos de la Universidad de Barcelona
- Agustín Cañas, Visual design for Learning Analytics. Do more with less, GRADIANT (Galician Research and Development Center in Advanced Telecommunications).
- Abstract. Modern education is changing the rules. Teachers and learners use multiple tools to support learning with technology. They generate large amount of data from their teaching and learning processes but, do they really analyze this data? do they really understand the data?We seek to marry learning analytics with real-world learning applications using a visual approach. From the study of the most-used visual languages and HCI concepts, we construct fluid user interfaces based on a single principle: do more with less. Teachers must be focused on teaching, learners must be focused on learning, data must be focused on helping. If they don’t understand the information, there is no information. Taking a close look at learning data we can show teachers and learners what they want to see. Thus, we are developing intuitive data-driven user interfaces using, among others, Modern-UI or Material Design principles from Microsoft and Google respectively.