EVALUATION OF UNIVERSITIES WEBSITES BASED ON USABILITY AND MACHINE LEARNING WEB CLUSTERING

Abstract

A university website is considered to be its main gateway to the world. The usability of a university website affects the way prospective students look for a program in a university for the first time, students who are using the website, and university staff who are accessing the website in their day-to-day tasks. That being said, there is a lack of research when it comes to evaluating the usability of universities’ websites using their web vitals. The purpose of this research is to evaluate the usability of universities’ websites using web vitals, and cluster universities based on the usability of universities’ websites. This research has three objectives. The first objective is to measure the response time of universities’ websites using web vitals to be used as a metric. The second objective is to examine the relationship between the usability of universities’ websites and university ranking systems. The third and final objective is to apply unsupervised machine learning for clustering universities based on the usability of their websites. The findings of this research will benefit universities when it comes to understanding more about the web vitals of their websites, examining the relationship between university ranking systems and the usability of universities’ websites, and exploring the most severe flaws that occur in universities websites. This research was carried out in three phases. Phase 1 started with measuring the performance, accessibility, best practices, search engine optimization, progressive web application, and other response time metrics, and verifying the collected data as well as collecting the universities’ ranking information. Phase 2 is where the data is cleaned and exploratory data analysis is applied to examine the relationship between the usability of universities websites and university ranking systems, and the flaws of universities websites were explored to determine the most severe flaws. Phase 3 is where the dataset was visualized and an unsupervised machine learning clustering algorithm was applied on the data to group universities based on the usability of their websites. This research has distinctively demonstrated that web vitals are effective metrics when it comes to precisely measuring the response time and the overall usability of universities’ websites, which addresses the identified research gap. Also, this research showed that the relationship between the usability of universities’ websites and the QS world university rankings 2023 and THE world university rankings 2023 is weak. In addition, this research grouped universities based on the usability of their websites into three different clusters.