With access to five years of course data, students create new apps at the Learning Analytics Hackathon

Over 50 students, researchers, and instructors attended UBC’s sixth learning analytics hackathon on November 2 and 3, spending the weekend creating apps and dashboards to improve student learning and experiences. Learning analytics is an emerging field that aims to use educational data – such as activity data generated in learning management systems like Canvas – to improve teaching and learning.

For the first time, participants were given access to 5 years’ worth of UBC’s course calendar data, which includes descriptions, pre/co-requisites, enrolment numbers, instructor names, and more, about every course taught at UBC from 2014-2018.

Combined with the data accessible via the Canvas API, participants worked with a richer set of data than at previous hackathons.

“The hackathon [participants] this year [are] really diverse,” said Stoo Sepp, a hackathon organizer and Manager of Learning Design in the Faculty of Education. To ensure that the participants could form productive teams, Sepp led an icebreaker activity to help them find others with a complementary skillset.

Doreen Mushi, a visiting scholar from Tanzania, found her hackathon partners during the icebreaker. “I have a very diverse group – it is pretty smart,” she said. Mushi, who is interested in learning design, worked with Kate and Mary, who are both undergraduates with good data skills, to build a tool that makes it easier to search for UBC courses.

Despite the limited time the teams had to come up with ideas and implement them, the quality of the projects was exceptional.

“For me, the best part of the hackathons are the student final presentations,” said Alison Myers, a veteran hackathon organizer and Research Analyst at the Sauder School of Business. “I was incredibly impressed with the ideas for the final projects, as well as how far students were able to get actually building tools that work.”

Team PI R Squares won first place for building an application to improve Canvas discussions. “Student discussion posts for courses can become very overwhelming, especially in bigger courses,” said team member Alexander Hinton. To make discussions more manageable by reducing duplicate posts, the team built an application that displays a list of pre-existing, relevant discussions in real time while a user types in a new discussion.

Runner-up winner BUCSE (Better UBC Course Search Engine) allows students to search for courses without relying on an exact keyword match. By using Google’s universal sentence encoder, the team was able to build a natural language processing model that computes a similarity score depending on how semantically similar the search term is to the course descriptions.

As team member and Master of Data Science student Braden Tam explains, while a search for “music” in the UBC course calendar returns many results, a search for “Mozart” comes up empty. Using BUCSE, courses teaching musical composition and musical literature are displayed.

The Learning Analytics project is exploring the feasibility of implementing some of these projects for campus-wide use. The organizers are also exploring ways of giving participants access to even more data at future hackathons. With more data, participants will be able to build even more innovative tools to support teaching and learning at UBC.

“The projects that students developed at the hackathon this fall were some of the best I’ve seen,” said Craig Thompson, a hackathon organizer and Research Analyst on the Learning Analytics project. “I’m really excited to see what new projects students will create at our next hackathon!”

Interested in taking part next time? Stay tuned for more info about the Learning Analytics Hackathon coming up this spring.

The Learning Analytics hackathon took place on November 2 and 3 in the Sauder Learning Labs, a space that’s designed to facilitate teams in using critical thinking and analytical decision making to solve real-world problems. The event was hosted by the UBC Learning Analytics project, LAVA (Learning Analytics, Visual Analytics), and Sauder Learning Services, with staff volunteers from Faculty of Arts, Faculty of Education, Sauder School of Business, and the Centre for Teaching, Learning and Technology.