Workshop Code Repository

https://github.com/nirmalpatel/edviz-2019

Workshop Schedule

The workshop will be split in two one and half hour sessions, split by a coffee break:

• Learning Analytics for Educators: In this part of the workshop, researchers will present their work around design and implementation of data visualizations that are useful for educators (among others) to make decisions.
  • Presentation 1 (25 min): Implications of Instructor Analytics Use Patterns for the Design of Actionable Educational Data Visualizations
  • Presentation 2 (25 min): "What do students know, how long does it take them to know?" at a Glance for Teachers and Instructional Designers
  • Interactive discussion (30 min): During the discussion, participants will be able to pose visualization challenges to others and ask them to quickly come up with data visualization ideas to communicate actionable information.
• Learning Analytics for Learning Designers: In this part of the workshop, researchers will present their work around data visualizations that can help learning designers improve on the educational material using data.
  • Presentation 1 (25 min): Visualizing the Solution Space of Educational Games using TRESTLE
  • Presentation 2 (25 min): Visualizing Cronbach’s Alpha for a Large Number of Assessments
  • Interactive discussion (30 min): During the discussion, participants will be able to pose visualization challenges to others, and ask them to quickly come up with data visualization ideas to communicate actionable information.

Introduction

We are pleased to announce a workshop on educational data visualization. One of the primary aims of this workshop is to create more value out of educational data by bridging the gap between educational data research and teaching practice. We are hoping to invite both researchers and practitioners and engage with them in a productive discussion on how to 'see' or visualize data in a way that guides decision making. The workshop will be organized around two topics:

1. Recommending evidence-based actions using educational data
2. Communicating learning analytics research to education stakeholders

Submissions will have to be focused on either of these two topics. The first topic aims to make learning analytics research more actionable, and the second topic aims to make learning analytics research more accessible. For further details on these two topics, please see the Submission Topics section below.

Our workshop puts a big emphasis on using open source software technology and reproducible research. All of the workshop submissions will be required to open source their data visualization programs written in either R, Python, or JavaScript or publish their visualizations in LearnSphere environment (either is fine.) Each code submission will be required to take one or more text files as input (plain text, CSV, TSV, JSON etc.) and produce the data visualization by using data from only those input files. It is highly encouraged that the final visualization is something that is of a publication quality and can be exported as an image, so that it can be included in academic papers. See Code Submission Guidelines for a requirements checklist.

At the end of the workshop, a GitHub repository combining all of the data visualizations will be made public. This repository will have programs to reproduce all of the data visualizations presented in the workshop.

Background

A recent survey in the United States found that 95% of the K-12 teachers use a combination of academic data and non-academic data to understand their students’ performance. However, 34% of the surveyed teachers also reported that there was too much data for them to look at. How can we help educators make sense of large amounts of student data? Data visualization is one of the most widely used techniques that help people make sense of large amounts of numerical information. Graphical representations of data can be used very effectively to communicate context-specific information.

Reporting of learner data is one of the cornerstones of LA research, and the LA community has developed domain-specific data visualizations to show student learning in different contexts. Some of these visualizations emerged from the Learning Science research community (e.g., Learning Curves,) while other visualizations have a close affinity with classroom practice (e.g., Curriculum Pacing Plots.) Although these visualizations are slowly making their way into the hands of educators, many of these visualizations are not easy to reproduce. For this reason, this workshop will also publish an open source gallery of education data visualizations that are easily reusable by LA researchers and practitioners.

Submission Topics

The workshop will be centered around two main topics:

1. Recommending evidence-based actions using educational data: How can we use educational data to recommend useful actions to educators and other education stakeholders? Submissions in this category will be focused on recommending actions using educational data visualizations. For example, a data visualization can suggest teachers which students need personalized interventions, or it might suggest an administrator a professional development plan, or it might show a MOOC course designer which alternate learning paths that students are taking. In summary, submissions in this category should focus on evidence-based decision making using educational data.
2. Communicating learning analytics research to education stakeholders: How do we use effectively communicate learning analytics research to a broad community of education stakeholders? Beyond academia, learning science and educational data research is valuable to people in many different ways, but how can we make it more accessible? For example, how do we make effect sizes more meaningful? How can we reveal the inner workings of recommendation engines?
   • Education involves many stakeholders and each stakeholder can benefit differently from LA research.
     • Educators might ask "How can I make decisions in my classroom using educational data research?"
     • Parents might say "Are there any research-based academic interventions that are useful to my kids?"
     • Administrators will ask "What will increase educational outcomes for my school/district?"
     • Policymakers will want to know "How can we be more scientific in making educational policy decisions?"
     • Textbook Publishers will say "How can we make our textbooks and digital programs better suited for learners?"
   • In summary, submissions in this theme will be focused on make LA research more accessible to a broad range of education stakeholders.

Each submission will have to fall into one of these two themes. If you think your submission does not fall into any of the above themes but is still relevant to the workshop, please send us an email and we will work it out.

Article Submission Guidelines

All papers must be original. We are only accepting short papers for this workshop. Each submission should:

• Be submitted before 3 December, 2018 23:59 PDT
• Be in LAK Companion Proceedings Format
  • Word template
  • LaTeX template is not available, please contact LAK committe if you need it
• Be no more than 6 pages long (excluding references)
• Be deanonymized (yes, deanonymized)
• Motivate the need behind the data visualization
• Describe the design space or design variations of the data visualization
• Show a few examples of the data visualization
• Have GitHub repository link to the data visualization code (please see Code Submission Guidelines)

If the data visualization was used in a field study, we highly encourage authors to include the results of the study.

Please submit your articles directly to Nirmal Patel (nirmal@playpowerlabs.com) and include "[EdViz-2019]" in your email subject.

Workshop proceedings will be published in the LAK Companion Proceedings.

Code Submission Guidelines

You do not have to submit code for your data visualization if it is published in LearnSphere environment.

Otherwise, all data visualization programs should:

• Be hosted on a public GitHub repository
• Be written in either R, Python, or JavaScript (we hate to see MATLAB not being included here, but we are committed to using open source technology and Octave is not widely used)
• Only ingest plain text files that are one or more of the following formats:
  • CSV file(s) (highly preferred)
  • Text file(s)
  • JSON file(s) (please consider using NDJSON if possible)
• Be able to save a PNG/JPEG image to disk

We highly encourage authors to make their visualizations publication quality.

Venue

The workshop will be held in conjunction with 2019 Learning Analytics and Knowledge (LAK) conference at Arizona State University.

Committee

• Nirmal Patel (Chief Data Scientist, Playpower Labs)
• Collin Sellman (Faculty Associate, Arizona State University)
• Derek Lomas (Assistant Professor, Delft University of Technology)

Contact

For any queries, contact Nirmal Patel (nirmal@playpowerlabs.com)