Previous Events

Open Science Research Practices for the Era of Big Data

Advances in several scientific fields has been facilitated by amassing large datasets and making sense of them by using cutting-edge statistical learning techniques. However, there is still one aspect that needs attention; embracing open science (OS) practices. The session will present how other fields are self-correcting by adopting OS practices, the OS-attached benefits of reproducibility and transparency, and the related issues of replicability and generalisability. The session aims at discussing opinions and avenues that can permit or hinder OS practices in the field of learning analytics and educational data mining. Attendees are required to have a genuine interest in the advancement of their field in order to jot down honest and viable ways to adopt OS practices. For this purpose, an online survey will be run and discussed with the attendees.

Organizers: Fernando Marmolejo-Ramos – University of South Australia

Target audience: Students, practitioners, and researchers in diverse areas but with interest in open science

Target level: Introductory and intermediate

Delivery date and time: 10 March 2021, 8.30am Australian Central Time/9 March 2021, 4pm US Central Time

Recording: https://www.youtube.com/watch?v=eSnheHE5P4U

Slides and other materials are available in the Resource Hub.

Connecting Data with Personalised Feedback

The abundance of data on how learners engage with a learning environment has increased significantly due to the ubiquitous presence of technology mediation. Instructors may use this data to increase their level of understanding of how learners are participating and make decisions about which actions would have the biggest impact in the overall quality of the experience. However, the connection between data and actions is not trivial and requires the alignment of artefacts, processes and methods and sometimes even fall within multiple organisational substructures. In this session we will explore a tool for instructors to articulate this connection more effectively and focus on support actions in the form of personalised feedback messages. Attendees are required to have experience on design and/or delivery of learning experiences and a sense of how to support learners throughout the experience.

Organizers: Abelardo Pardo – University of South Australia

Target audience: Academic developers, designers, instructors. Anybody with experience designing or deploying learning experiences that require student support

Target level: Intermediate. Some knowledge about data available in institutions and knowledge about how to support students is required.

Delivery date and time: 23 February 2021, 7:30am Australian Central Time/22 February 2021, 3:00pm US Central Time

Resources:

Recording: https://www.youtube.com/watch?v=ZowVG8iAxRA&feature=youtu.be

6 Techniques for Participatory Design of Learning Analytics with Students and Teachers

This session introduces learning analytics researchers and practitioners to a collection of tools that can be used to scaffold the effective and authentic participation of stakeholders in the design, implementation and evaluation of Learning Analytics systems. The session is practice-based. After establishing the power of human-centered learning analytics design to create systems that resonate with stakeholders’ needs, we will share six tools and techniques that can be used to Understand, Co-Design, and Evaluate them. As part of the session, participants will work on sketching out the first stages of a project for the human-centered design of learning analytics.

Organizers: Juan Pablo Sarmiento – NYU Steinhardt/LEARN

Target audience: Learning analytics professionals interested in human-centered design, faculty interested in Learning Analytics innovation, and instructional support professionals at educational institutions

Target level: Introductory

Delivery date and time: 10 February 2021, 4pm US ET)

Resources:

Recording: https://www.youtube.com/watch?v=fvlIWgKAeSA&feature=youtu.be

 

Tools and Experiences in Learning Analytics

We invite you to come join us for a seminar that will focus on adoption and use of learning analytics in K-12 schools and post-secondary institutions in Europe and Latin America. More information about the talks by each speaker is available below.

  • 15:00 – 15:40 – Learning Analytics for Large Scale Data – Alexandra Cristea, Professor, Head of the Innovative Computing Research Group, and Deputy Head in the Computer Science Department at Durham University.
    • Recent learning analytics approaches in different learning environments (MOOCs, and online and blended courses) by the Artificial Intelligence and Human Systems research group.
  • 15:40 – 16:20 – Applying Learning Analytics in Living Labs for Educational InnovationTobias Ley, Professor of Learning Analytics and Educational Innovation and Head of Center of Excellence on Educational Innovation at Tallinn University.
    • Using Living Labs for educational innovation centered on the theory-guided and stakeholder-driven application of learning analytics in secondary schools.
  • 16:20 – 17:00 – Adaptation, Adoption and Learning Analytics Pilots in Latin America – Pedro J. Muñoz Merino, professor at Universidad Carlos III de Madrid and LALA’s coordinator.
    • Presentation on the LALA Framework, tools for counseling, prediction for drop outs, and discussion of different tools used by students and faculty.

Organizers: eMadrid (Co-Sponsored by LALN and UNESCO Chair on Scalable Digital Education for All

Target audience: All, including K-12 and Higher Education practitioners, administrators, instructional support staff, and researchers

Target level: Introductory

Delivery date and time: 15 January 2021, 3:00-5:00 local time (Central European Time)/8:00-10:00am US CT

Recordings and Slides:

Learning Analytics, Equity, and Social Change (Part II)

In this conversation series, participants will continue the discussion and ideas around how to build equity and social change within the different communities and learning spaces within the field of Learning Analytics. The facilitators will use SoLAR’s Statement of Support and Call for Action on Social Justice & Dismantling Injustice in Education as a starting point to reflect on participants’ local contexts, institutions, and communities. The group will engage in activities and open discussion centered on Learning Analytics research and articles to dissect how the insights, questions, or findings can be applicable to each participant’s context. This session is an open the conversation where students, staff, and faculty are encouraged to engage with the discussion to explore the importance of equity and social change, and how these topics can enhance the field of Learning Analytics in order to better educational systems.

Pre-Event Readings (Optional!):
SoLAR Executive Committee: Statement of Support and Call for Action
Should predictive models of student outcome be “colour-blind”?
What’s the problem with Learning Analytics?
“We called that a behavior”: The making of institutional data

The first conversation included a presentation on participatory, co-developed, learning experience within the Learning, Design, and Technology Program at Georgetown University from both the student and faculty perspectives to prompt discussion around the question: How do we embed equity and social change within the curricula of Learning Analytics?

Organizers: Aaron Joya and Yianna Vovides, Georgetown University

Target audience: Students, Practitioners, Researchers

Target level: Introductory

Delivery date and time: 17 December 2020, 4:00pm – 5:00pm US CT

Recording: https://georgetown.hosted.panopto.com/Panopto/Pages/Viewer.aspx?id=e23af70b-ecad-4a84-bc0b-ac94017eeeee&start=2.236505

An Introduction to Writing Analytics and the Design of Automated Feedback

Writing Analytics is a subfield of Learning Analytics that focuses on the challenges learners face in writing. It uses natural language processing (NLP) and machine learning technologies to analyse texts, and can be used to provide automated formative feedback on student writing.

This online workshop introduces educators and researchers to the affordances of writing analytics and automated writing feedback tools for students. We will first provide an overview of writing analytics techniques with examples of their educational applications. We will then demo and explain the usage of the open-source tool ‘AcaWriter’ developed by the Connected Intelligence Centre, University of Technology Sydney, which provides all students 24/7 instant feedback on their writing. We will provide practical context by explaining how we co-design and evaluate this tool with educators and students, and how it is integrated with learning design.

A set of resources including Python code to generate sample automated feedback messages for technical participants, a learning design template for educators, and follow-up readings for all, will be shared with workshop participants to enable them to go deeper after the event.

Organizers: Antonette Shibani and Simon Buckingham Shum, University of Technology Sydney; Ming Liu, Southwest University, China

Target audience: Practitioners, Researchers

Target level: Introductory

Delivery date and time:  20 November 20, 2020,  11:00am AEST (19 Nov, 6:00pm US CT)

Recording: https://www.youtube.com/watch?v=hMJODVc9yfk 

Slides and other materials are available in the Resource Hub. More information and additional resources are available at http://wa.utscic.edu.au/events/laln-2020-workshop/

Learning Analytics, Equity, and Social Change (Part I)

In this conversation series, we hope to stimulate discussion and ideas to engage with topics of equity and social change within different communities and learning environments in the field of Learning Analytics.

In this first conversation, we will begin by presenting on a participatory, co-developed, learning experience within the Learning, Design, and Technology Program at Georgetown University from both the student and faculty perspectives. This initial sharing will prompt and open the conversation to address the following question: How do we embed equity and social change within the curricula of Learning Analytics?

The second conversation in this series (December, 17 4pm US CT) aims to engage those within and outside higher education in further exploring how equity and social change can inform the field of Learning Analytics.

Organizers: Aaron Joya and Yianna Vovides, Georgetown University

Target audience: Students, Practitioners, Researchers

Target level: Introductory

Delivery date and time: 12 November 2020, 4:00pm – 5:00pm US CT (Part II – December 17, 4-5pm US CT)

Recording: https://georgetown.hosted.panopto.com/Panopto/Pages/Viewer.aspx?id=2efd42b7-3de4-4ea5-b68e-ac710186602e

Hyperparameter Tuning in Machine Learning for Student Models

Hyperparameters are the parts of a machine learning model that are not directly learned from student data. Things like which features to put into the model, how complex the model should be, and even what algorithm to use are all hyperparameters that have a huge impact on results, and must be chosen carefully. In this tutorial, we will go over different strategies for selecting hyperparameters, for models based on logistic regression to random forests and deep neural networks. We will also highlight common pitfalls and challenges in hyperparameter selection, and point to various tools that can help learning analytics researchers implement hyperparameter tuning quickly and correctly.

Organizers: Nigel Bosch – University of Illinois

Target audience: Researchers

Target level: Intermediate (some machine learning experience)

Delivery date and time:  20 October 2020, 4:00pm – 5:30pm US CT

Recording: https://www.youtube.com/watch?v=YSMc69e-smY 

Slides and other materials are available in the Resource Hub

Train an AI Model with BERT; or, Everything I Know about NLU I Learned from Pop Culture

The 2020s have seen an explosion of pre-trained language models, from BERT and Ernie to GPT-3 and others. These emerging models can serve to jump-start natural language processing (NLP) and possibly natural language understanding (NLU) in a variety of areas, but what value do they potentially bring to educational researchers working with qualitative or “thick” data?

In this workshop, the opening presentation will provide theoretical background and examples of pre-trained LMs for potential use by researchers in educational settings. Then, participants will utilize a publicly-available AI modeling platform to analyze sample texts, with a guest appearance by one or more American pop culture figures. A short bibliography of follow-up readings will also be provided.

Organizers: Pete Smith, Henry Anderson, Elizabeth Powers, Justin T. Dellinger – University of Texas at Arlington

Target audience: Practitioners, Researchers

Target level: Introductory

Delivery date and time:  18 September 2020 3:00 pm – 4:30 pm (US Central Time)

Recording: https://www.youtube.com/watch?v=JB90xiA8LoQ

Slides and other materials are available in the Resource Hub

Trusted Learning Analytics

Learning Analytics promises the chance to improve learning and teaching. By analysing learning process data, learners and teachers could be provided with information about their learning process. This data might support the learners’ self regulation, improve the assistance by teachers, or might be used for automatically generated support like interventions. However, the risks shouldn’t be underestimated. Permanent monitoring of students is often hindering them, e. g. to ask questions or to speak freely, in particular if the data is stored for later usage. To learn effectively, feedback of half-baked thoughts are necessary. Therefore, trust is required in all parties concerned, including the utilised platforms and technology. Any abusive data usage, like using data for other purposes as stated, must be ruled out.

The innovation forum Trusted Learning Analytics researches various possibilities to decrease these reservations, and to support the distribution of Learning Analytics in universities by using model-like steps. In this workshop we would like to discuss the aspects of learning analytics in general, but also want to focus on international aspects and differences.

Organizers: Oliver Herrmann – Goethe University Frankfurt am Main

Target audience: Practitioners, Researchers

Target level: Introductory

Delivery date and time:  14 September 2020 9:00 am (US Central Time) / 16:00 local time

Recording: https://www.youtube.com/watch?v=4txLnqtKIkU 

Other resources are available in the Resource Hub

Feature Engineering: Better, More Interpretable Models

This tutorial will discuss how to distill and engineer features for data mining, and the advantages and disadvantages compared to automated methods such as auto-encoders. We will cover the process of feature engineering and distillation, including brainstorming features, deciding what features to create, and iterative feature engineering. We will discuss what tools are particularly useful for feature engineering.

Organizers: Ryan Baker, Stefan Slater – University of Pennsylvania

Target audience: Practitioners, Researchers

Target level: Introductory

Delivery date and time: 24 August 2020 3:00 pm – 4:30 pm (US Central Time)

Recording: https://www.youtube.com/watch?v=ko5WEfLqCh8

Slides and other materials are available in the Resource Hub

LAK Workshop – Building Capacity Through the Learning Analytics Learning Network

Organized by Justin T. Dellinger, Florence Gabriel, Ryan Baker, George Siemens, Shane Dawson – University of Texas at Arlington, University of Pennsylvania, and University of South Australia

Event canceled due to COVID-19. More information available at https://sites.google.com/view/laln/home/lak20-workshop and https://lak20.solaresearch.org/pre-conference-details

Learning Design for Learning Analytics

Organized by Yianna Vovides – Georgetown University

More information is available at https://learninganalytics.net/laln/2020/03/18/laln-dc-chapter-event-march-23/

NLU = NLP + AI ? Imagining the Future of Language and Culture Computing for Education

The 2010s have been labeled the “decade of NLP”—important progress was made in computational approaches to language data and tasks such as thematic analysis, text classification and clustering, as well as machine translation. But what does the next decade hold for researchers and practitioners with qualitative or “thick” data—and does it look more like Natural Language Understanding than Natural Language Processing?

In this workshop, the opening presentation will provide definitions and examples of NLP for educators, review emerging definitions of NLU, and prepare the participants to utilize one major NLU tool with educational text data, in the context of discussing deeper understanding of text and what this may entail. Current, ongoing debates on the national stage about the future of AI will also inform our views of the future. A short bibliography of follow-up readings will also be provided.

Organizers: Pete Smith, Henry Anderson, Elizabeth Powers, Justin T. Dellinger – University of Texas at Arlington

Target audience: Practitioners, Researchers

Target level: Introductory

Delivery date and time: 28 February 2020 3:00 pm – 4:30 pm (US Central Time)

Recording: https://www.youtube.com/watch?v=Shk5eIg-aoE&t

Materials: All session materials, including references, bibliography, and facilitator’s guide, are available in the Resource Hub

Introduction to Learning Analytics

With the increased adoption of technology, institutions have unprecedented opportunities to continuously improve the quality of their services through data collection and analysis. Schools and universities now have data about learners and their contexts that can provide valuable insight into how they learn. Early attempts were directed towards mining educational data to identify students-at-risk and develop interventions. Recently, more sophisticated approaches are being deployed by researchers and practitioners. These include analysis of learner behaviour that leads to various learning outcomes, social networks and teams, employability, creativity, and critical thinking. Analysing digital traces generated through learning processes requires a broad suite of methods from data science, statistics, psychometrics, social and learning sciences.

This workshop aims to introduce teachers and educators to the fast growing and promising field of learning analytics. How digital data can be used for the analysis and improvement of student learning will be explored. First, we will provide an overview of learning analytics, its key methods and approaches, as well as problems for which it can be used. Secondly, attendees will engage in group learning activities to explore ways in which learning analytics could be used within their institutions. The focus will be on identifying learning-related challenges that are relevant to their particular context and exploring how learning analytics can be used to practically and effectively.

Organizers: Vitomir Kovanovic and Srecko Joksimovic – University of South Australia

Target audience: High school teachers and management

Target level: Introductory

Delivery date and time: 22 October 2019 10 am – 12 pm (ACDT)

Recording: https://www.youtube.com/watch?v=VB9nXkcLXW4&t=21s

Slides: https://www.slideshare.net/vitomirkovanovic/introduction-to-learning-analytics-for-high-school-teachers-and-managers