In this week’s meeting Leah Macfadyen and Alison will be sharing work that we’ve dubbed “The IKEA Model”.
Initially developed as a Directed Study in the MET program, the goal was to build an instructor-facing learning analytics dashboard using data that is accessible from Canvas by an instructor. The “project” includes a series of steps and scripts that would allow you to pull together Canvas and New Analytics data into a Tableau dashboard. We will discuss the background of the project, our roles, and its current state.
An essential part of analytics is the communication of insights to colleagues and stakeholders. In this LAVA session, we will look at Quarto, an open-source scientific and technical publishing system that combines code and its outputs with descriptive text and exports them to many formats, including HTML, PDF and MS Word. We will cover standard tools to author Quarto documents, how to format them using Markdown, and how to include code with R as an example (Python and Julia are also supported).
Preparation (entirely optional):
We will briefly cover Markdown but will only have time to go over some things in the basic syntax. If you would like to familiarize yourself with Markdown beforehand, here is an excellent interactive tutorial (10–20 min): https://commonmark.org/help/tutorial/
If you would like to work along the session to create a Quarto document, you have two options:
Stephan previously worked as a Postdoctoral Teaching and Learning Fellow in the Microbiology and Immunology Department at UBC and coordinated the Experiential Data science for Undergraduate Cross-disciplinary Education (EDUCE) program. He designed and delivered data science-focused modules across multiple MICB courses, conducted and analyzed surveys, and maintained open education resources. He joined Skylight (Science Centre of Learning and Teaching) as the Science Education Specialist in the Computer Science Department and supports instructors with the migration to a new online assessment tool offering students the opportunity to practice with randomized questions for mastery learning and more flexibility when scheduling exams in a new computer-based testing facility. He focuses on the impact of learning technologies on student learning, well-being and inclusivity and how their implementation and course design should inform each other.
In this week’s LAVA meeting Alison shared how Sauder Learning Services takes advantage of Canvas APIs. Alison is a Research Analyst in the Sauder Learning Services team. The team supports our faculty’s use of technology in their teaching practice. In her role, she attempts to understand what data is available and can be used to inform teaching practices.
In this LAVA talk I will be discussing how my team uses available Canvas APIs for completing routine tasks and extracting data. I will briefly share what we have learned about the differences between REST and GraphQL APIs at a high level (as the technical details are beyond me). I will also share some of the projects that we have used the APIs for, and discuss our approach to the development of these projects from an operations standpoint (how we try to “reduce clicks” and save time), and from a learning analytics standpoint (how to use the data available to inform teaching practices).
In our August 4th LAVA meeting, we will be hearing from Dr. Fatemeh Salehian. Fatemeh is a postdoctoral research fellow in the School of Information at UBC. She is currently affiliated to the University of Michigan, Ann Arbor, and the University of South Australia. She has been working on learning analytics projects for over seven years. Her research focuses on understanding how learners self-regulate their learning processes and how learning analytics can help to support self-regulated learning since it is critically associated with students’ performance and learning outcomes. Her research focuses on using data to study learners’ behaviors in online learning environments to develop indicators of learning behavior.
I will talk about two projects focusing on providing actionable information for students.
First is the OnTask platform which offers an intuitive interface to upload data about student engagement into a matrix and the definition of a set of simple when/do rules to customize email messages for the students. Instructors use the indicators about student engagement to select/ignore text portions that are then collated and sent to the students as regular email messages. These messages provide students with personalized support and suggestions to improve their engagement. In our recent analysis, the results showed that there is a significant association between the messages’ topics and the students’ performance.
Second, My Learning Analytics (MyLA) is a student-facing dashboard that provides students with information about their engagement with course materials and resources, assignments, and grades in a Canvas course. A set of three learning analytics data visualizations have been designed to reveal behavioral patterns associated with good learning skills, guide decisions about actions students can take that may improve their academic outcomes, and provide a transparent view of students’ course standing.
This week we will be hearing from Dr. Bowen Hui. Dr. Hui is Associate Professor of Teaching & Associate Head of Undergraduate Affairs (CMPS) in the Computer Science program at UBCO.
Teamable Analytics: A Team Formation and Analytics Tool
Forming effective teams for large classes is a challenge for educators due to the complexity of project needs, the diversity of individual characteristics, and the criteria different educators have for forming teams. Although many researchers over the past several decades studied the success factors of a team, there is still little consensus on how a team should ideally be formed. Consequently, how one decides to form teams in a class depends on the domain, classroom context, and pedagogical objectives.
In this demo, we present a web application that offers several algorithms to support the team formation process called Teamable Analytics. Teamable Analytics is compatible with any learning management system (LMS) that uses the LTI protocol. Our tool provides a dashboard for educators to elicit student characteristics and customize how those responses are combined to form teams. In contrast to existing team formation software, our tool supports more use cases for building teams, it provides a general end-to-end solution to the team formation process, it is integrated with the LMS so to minimize data setup and privacy risks, and it makes use of visual analytics to diagnose problematic formations, increase user trust, and monitor ongoing team performance.
To date, our team has integrated Teamable Analytics with the Canvas LMS and completed 7 pilot studies with interdisciplinary classes consisting of 15 to 210 students across both UBC campuses. We are currently seeking collaborators to further the research and development in the visual analytics component of the project.
This week’s LAVA session [Thursday April 7, 2-3pm, Zoom] will be led by Annay Slabikowska (PAIR) and Craig Thompson (CTLT), who will present recent work on the Student Flows Project.
From Annay and Craig:
The project is collaboration between the Learning Analytics and Planning and Institutional Research teams and the Faculty of Science, including contributions from analysts in the Faculty of Arts and the Sauder School of Business.
The goals of the Student Flows Project are threefold: to create a shared common data set that can be used by analysts across the institution to explore student flow scenarios, to develop visualizations of student flows and to recommend tools to visually analyze student flows, and to collaboratively document and share best practices for exploring student flow scenarios.
By “Student Flows” we mean student progression through courses, specializations, and degree programs during their time at UBC. This can include analysis of time to degree completion, identifying common patterns of switching between faculties or majors, investigation of common course co-enrolment or sequencing, and more. In addition to these enrolment-based data elements, we are also interested in student characteristics, and how they relate to enrolments. For example, we need to be able to explore whether there are differential course or program outcomes for various student cohorts, such as international students, mature students, or indigenous students, as well as temporal cohorts.
From our initial exploration of this space, it became clear that there was not a one-size-fits-all solution to data analysis and reporting regarding student flows. For example, the level of detail needed varies by audience: one Dean may be interested in understanding how students flow into and out of their Faculty, whereas a department head may be interested in how students choose to sequence courses within their major. Additionally, at any given level of detail, different stakeholders will have different needs terms of what information they want to see, and how they would like it to be presented. Thus, we have endeavored to create a framework (dataset, documentation, and established best practices) that will enable a distributed team of analysts to undertake work in this space, to more easily collaborate and share, and to answer the plethora of questions raised by various institutional stakeholders.
We draw inspiration from other universities such as University of Michigan, which is an early leader in establishing a common data set for institutional researchers, and numerous institutions exploring visualization techniques for student flows such as UC Davis, University at Buffalo, and University of New Mexico.
For our first LAVA meeting this year (Feb 3, 2-3pm) Alison Myers (me!) will be doing an introduction to Tableau, with a short demonstration of some of the dashboards that Sauder Learning Services has developed over the past few years (many of you will likely have seen these dashboards in some form from past sessions).
This will be a speed run through of Tableau to introduce basic concepts, while briefly touching on some more intermediate functionality. If you are unfamiliar with Tableau – this will likely be “too fast” – but will hopefully give you a sense of the possibilities. If you are familiar with Tableau, then hopefully I might introduce a tip or trick you haven’t seen. If anyone has anything specific that they would like demonstrated or discussed – please feel free to reach out and let me know ahead of the session.
Join us on Wednesday July 07th at 3pm-4pm PST for our next LAVA session held by Dr. Julie Wei form the Faculty of Arts.
Reframing Student Evaluation of Teaching (SEoT): Using Student Comments to Unlock Hidden Themes
Student evaluation of teaching (SEoT) has been implemented for decades by many universities including UBC. This has created and is still creating huge amount of student feedback comments that pour in course and teaching evaluation survey, however this type of data has not been used widely, probably because it is time-consuming to review the unstructured data. In this presentation, Dr. Julie Wei will share a project she initiated and led at the Faculty of Arts by employing both students’ original comments and the explicit suggestions automatically extracted from them by using Natural Language Processing (NLP) approach. This could quickly provide instructors and decision-makers with useful information that could help further identify the areas that need to improve and thus help promote quality teaching and student success in the long run.
These sessions are being held via Zoom. To attend this session, or to be added to our mailing list to receive information about future sessions, please contact Alison Myers (firstname.lastname@example.org).
As many of you will know, the 11th International Learning Analytics and Knowledge Conference was held last week. So, this week at LAVA we’ll have an informal meet up and share highlights from the conference. If you attended LAK, it would be great if you could come prepared to discuss what you felt were highlights from the conference with the group (links, papers, insights, etc.)
Marko Prodanovic from Sauder Learning Services will be presenting their work on accessing and making use of Panopto video data.
Over the past year we’ve seen a sharp increase in teaching and learning happening through asynchronous video lectures. This also means a lot of new data!
In this session, Marko will show some of the work he’s done in developing a Python tool for fetching and managing viewing data from the Panopto video platform. We’ll look at the means by which data is being accessed (through various APIs), the transformations we perform and the reasoning behind them. We’ll also walk through a set of video analytics dashboards, built in Tableau that utilizes this data. As this project is ongoing, the last 15 minutes of the session will be dedicated to feedback, suggestions and discussion around the broader applications of this kind of viewing data.