Missed the activities fair? Don’t panic—there’s now an app to help you decide which organizations to join.
The Duke CoCurricular E-advisor—developed by Duke undergraduates participating in the Data+ summer research program—generates a ranked list of activities that fit the specific user. The e-advisor algorithm is still in its testing stage.
“The plan is to have some system where we can have new students put their general interests into the program, and from those general interests be able to give them recommendations,” sophomore Alec Ashforth, a member of the Data+ team, said in a video.
Students start by entering a Net ID, major, expected graduation date and—for upperclassmen—Duke activities they have participated in. After submitting a profile on the site and hitting the “recommend” button, the user is presented with a personalized list of organizations at Duke recommended by the e-advisor.
“You use actual student data of all the programs that they’ve participated in, and you feed that back into the system to help match their profiles to a new student’s profile,” Brooke Keene, a junior on the research team, said in the video. “The system will only improve if we can keep having students to test our system and keep inputting their data.”
New students can also indicate general interests from a set of keywords instead of inputting activities they’ve participated in.
The project, though still being developed, currently includes 150 student organizations and data from 80 students. Students can also add student organizations they are a part of to the database. The database only includes co-curricular programs at Duke that don’t count for credit.
“Co-curricular programming refers to all the learning that happens outside of your academic classes,” said Michael Faber, manager of the Innovation Co-Lab at Duke, in the video. “All the clubs, all the projects, independent work—things that are still considered learning opportunities.”
The team built the app using R Shiny, a framework for building web applications.
The team hopes to expand the application into a comprehensive product that can act like a comprehensive virtual activities fair.
“We built a recommendation algorithm using user-to-user matrices and user-to-item matrices, which is collaborative filtering and content-based filtering,” junior Dezmanique Martin, one of the researchers, said.
Get The Chronicle straight to your inbox
Signup for our editorially curated, weekly newsletter. Cancel at any time.