As a graduate student, I know the process of looking into which school and program one should apply to is a long and arduous one. For example, you probably don’t know the ideal GRE score needed to be accepted into your dream program, or you don’t know which schools offer the program you want to because the information available is scattered. To better deal with this complicated and disorganized process, a couple of friends and I built GradSeer, a graduate application assessment website that helps prospective applicants decide which graduate program and/or school they should apply to.

We knew that graduate admission data is not publicly accessible, so we instead looked at Gradcafe, a public forum with almost 500k data points of graduate admission results across the U.S. (for most graduate programs except M.B.A), submitted by prospective students after they were accepted or rejected into a program. Since we wanted to look at the average credentials of people who got accepted and rejected, we think Gradcafe was quite representative.

To collate information, we built a scraper to get data from the Gradcafe website, cleaned the data, and did various data visualization that you can see on the homepage. We also built a predictive analytic tool based on three Models (Logistic Regression, Support Vectors Machine, and K-Nearest Neighbors) that gives prospective students the likelihood of getting accepted to the program based on credentials. And finally, we set up a Google Cloud instance to build and store the web application.

Here’s how to use the predictive analytics tool.

1. Go to the Predictive Analytics menu-> Predict my Fate.

2. Say you want to apply to Columbia University Computer Science Graduate Program right after finishing your undergraduate program. You’ll need to fill in the fields below with your credentials. Note that GPA refers to your undergraduate GPA, and College and Major are your desired graduate program (all fields are mandatory).

3. Once your information is filled out, click “Generate Prediction.”

4. The results will give you three probabilities based on three different models. The result below means you have a 63.08% chance of getting into Computer Science at Columbia University.

5. Now, this is where it gets interesting. If you scroll further down the page, the algorithm will show you all the schools where you have a high chance of acceptance.

6. And finally, if you just want to see average statistics of each university or program, you can go to the Admission Statistic menu -> General Admission Statistics.

We hope this program will be useful in at least providing you with some insights into graduate admission statistics in the U.S. Finally, good luck with your application!

Mahendra Vitrianto, is a software engineering intern at MPOWER Financing from Columbia University. He is a clean code connoisseur, microservices architecture aficionado, and Python magician.

© MPOWER Financing, Public Benefit Corporation
NMLS ID #1233542

DISCLAIMER - Subject to credit approval, loans are made by Bank of Lake Mills or MPOWER Financing, PBC. Bank of Lake Mills does not have an ownership interest in MPOWER Financing. Neither MPOWER Financing nor Bank of Lake Mills is affiliated with the school you attended or are attending. Bank of Lake Mills is Member FDIC. None of the information contained in this website constitutes a recommendation, solicitation or offer by MPOWER Financing or its affiliates to buy or sell any securities or other financial instruments or other assets or provide any investment advice or service. © MPOWER Financing, Public Benefit Corporation 2019 NMLS ID #1233542. 1101 Connecticut Ave NW Suite 900, Washington, DC 20036