PhD Program Recommendation System: A Shiny App for Matching Applicants to University Requirements.

Abstract

The rising demand for graduate programs has created a need for a more efficient system to help potential students choose the right universities. This study presents the development of a PhD Program Recommendation System using R Shiny. The novelty of this system lies in its interactive design that offers real-time, personalized PhD recommendations using data-driven methods. The system is based on a dataset with the requirements for graduate admissions into USA universities, such as GRE, English test scores, and indicators related to research focus. Its user interface enables candidates to enter their academic profile details including GRE score, English tests (TOEFL, IELTS, etc), GPA and research experience. The app then dynamically filters universities that fit the qualifications of the candidate based on this input. We have used a cosine similarity algorithm to check the gap between what the candidate enters on his/her profile and university data. Each vector is normalized with L2 norm scaling to produce a better similarity score than raw counts, which is especially important when comparing universities across differing scales. The key result of the application includes a list of matched universities ranked in order, accompanied by details on required scores and research interests based on the similarity score. Second, there is a bar plot visualization showing comparative similarities of the candidate to the selected universities for better decision-making. The system is dynamic, in that as the user provides input to the system, recommendations change in real time. It will bridge the gap between students and universities by recommending appropriate PhD admissions with data-driven insights. The automation reduces manual search complexity, mitigates the risk of unsuccessful applications, and helps avoid unnecessary expenses by targeting programs that best align with the candidate's profile. While the system works well, it is limited by the available data and may not cover all aspects of university admissions. Future updates will expand the dataset and incorporate user feedback to improve the recommendations. This tool serves as a valuable resource for prospective PhD candidates, providing personalized, data-driven insights to navigate the highly competitive landscape of securing PhD opportunities.