Using application generated data to provide personalized user experience in software applications

Abstract

Delivering quality user experience is the most critical goal of any product development and marketing team in any organization. With the advancement of technologies in the fields of data science and data computation, it is now possible to know users more efficiently and create solutions that satisfy their needs to the fullest. In this thesis, I explore how the digital e-commerce and online content provider companies are utilizing many different personalization methods which are helpful in increasing the rate of successful transactions, however, a similar trend is not visible in SaaS applications. Cloud computation has made software both easily accessible and replaceable, putting a lot of stress on both the value of the product as well as the user experience. Many software companies still follow the traditional approach of creating static personas for product design and marketing purposes and create one fits all solution. Machine/application data, which is continuously generated by the software applications, tracking each and every user activity, can be extremely useful in understanding the user behavior and thus giving companies the ability to create more personalized and adaptive solutions. I explore data generated about a pedagogical website at MIT which is used to support instruction in computation-open to students from all the departments. I applied machine learning algorithms to show that there are different clusters/classes of students in a class. By tracking student activity and performance on class website, it can be predicted which class they belong to. This information can be used to develop customized solutions for all the students.