Abstract:
Performance issues in loading web page contents and developing them such way that
these becomes start interacting with users in fastest possible time has been a massive
challenge for businesses over the years. Any medium to large sized web application
may have several web pages and each of them may have several performance issues.
One can fix them all but in perspective of meeting deadlines and allocating budget it
can be cumbersome. The main objective of this research is to measure the impact of
those performance issues that are creating massive impact in terms of other websites
based on visual progression and developing a Continuous Integration, Continuous
Development tool as a part of GitHub Actions marketplace to recommend developers
and companies on prioritizing issues across various pages within their public and
private repositories. This project based research follows quantitative research
methodology including data collection, exploratory analysis of data, selection of
features in uni-variate and wrapper techniques as well as re-sampling through traintest-split and K-fold cross validation techniques in several regression algorithms. We
have developed a REST API by using dotnet framework to serve it as a service.
Finally, we developed a GitHub Actions by using JavaScript on Node.js runtime to
preview predictions on prioritization in several pages like a recommendation tool.
This research shows time to interactive, boot-up time and largest contentful paint are
the most important metrics in terms of boosting a site’s visual progression based
experience. This research is expandable in terms of other user experiences as well as
the project in terms of authorization mechanism.
