Ross School of Business and College of Engineering seniors Develop ML Platform to Predict Likelihood of Getting Into Classes

 

  • Ross School of Business and College of Engineering seniors create platform for students to predict chances of getting into classes at their registration time
  • Platform uses machine learning models to calculate risk of getting less-desired class sections, times
  • More than 3,000 students have signed up to use their website and performed over 35,000 predictions

 

March 30th, 2018 – From incoming freshman to students interested in taking in-demand courses such as EECS 183, there seems to be one commonality: the process of registering for classes can be as stressful and time consuming as actual coursework.

 

Ross School of Business senior Jordan Katz and College of Engineering senior Tyler Laredo came up with a solution to help students better understand what courses they’ll have a chance of getting into the following semester.

 

On their website, once students at the University of Michigan enter their assigned registration date and time, they simply type in the class they wish to register for and are given a percent chance of whether they’ll be able to get a seat in that course.

 

How did they do it? They began by tracking in real-time enrollment and registration data from the University. From there, they built a database comprising of more than 11 million snapshots of course enrollment statistics throughout the registration process. Then, using a machine learning model, they were then able to calculate the chance a student had of getting into a class section based off of their registration time.

 

They first launched their website just under a year ago, during the Fall 2017 class registration period.

 

Instantly, they saw a huge success—within two weeks, over 3,000 students had registered accounts on the platform.

 

“We were blown away,” said Katz. “Word of mouth was huge—people were telling everyone.”

 

For the most part, Katz said their website targets disadvantaged underclassmen who typically have the latest registration blocks.   

 

Now, one year later after their launch, they’re seeing users come back to sign-in by the thousands and many new first-time users as well as the Fall 2018 registration date slowly approaches.

 

“If people are using it and they’re signing on several times over the process, that means that it’s providing real value for people and that we’ve created something is making people’s lives easier,” said Katz.

 

Laredo, Co-founder and front-end developer of ClassAI, was a member of the first cohort of students in the Entrepreneurs Leadership Program back in Winter 2016.

 

With graduation on the horizon, Laredo and Katz have big plans for their futures. Come the fall, both will be headed to the West Coast to pursue careers in tech. Laredo has accepted a full time offer with Lyft as a Software Engineer and Katz will be working full time as a Product Manager at Asana.