Similarities in Challenges faced by Developers: Investigations on Stack Overflow and GitHub

Abstract

A large amount of rich data available in today�s world encounters a lot of opportunities to analyze the data and identify some useful patterns from them. However, dealing with such data requires automated frameworks and knowledge of programming languages. Java and Python are the most commonly and widely used programming languages among the developers which it is evident from the queries and issues posted on Stack Overflow and GitHub. Despite the popularity of both Java and Python, the challenges in transitioning from one technology to another technology are hard for individuals and industries. In this thesis work, we aim to investigate similarities in the challenges faced by the developers while dealing with both the programming languages. To achieve this goal, we formulated three research questions (RQs) for understanding the topics and issues asked and faced by developers. To achieve the results, we have used the topic modeling technique Latent Dirichlet Allocation (LDA).We have also identified the temporal trend of asking new questions on Stack Overflow for Java and Python programming languages (PLs). Our results revealed the changing trend, in the year 2015 onwards, from Java to Python and inclined towards Python from the number of the new posts in a Stack Overflow. We performed analysis on 18,892 Stack Overflow uestions related to Java and Python PLs and 42,674 issues from 22 different GitHub repositories, 11 for each PL. Our results indicates that questions asked on the Stack Overflow are co-related to issues posted by developers on GitHub during real-time development for a respective PL.