Analyzing Developer-ChatGPT Conversations for Software Refactoring: An Exploratory Study

In recent years, Large Language Models (LLMs) have witnessed a remarkable ascent, with OpenAI’s ChatGPT, introduced in 2022, garnering substantial attention. ChatGPT’s rapid adoption in the software development community has opened up new avenues for exploring its qualitative and quantitative impact on Developer-ChatGPT conversations. In this paper, we delve into a rich dataset from GitHub and Hacker News to perform a thorough analysis. Our objectives include characterizing the nature of these interactions and evaluating the use of ChatGPT in refactoring. To achieve these goals, we employ a combination of exploratory data analysis and data annotation, utilizing relevant keyword filters to extract pertinent information. Our examination encompasses the identification and analysis of code refactorings facilitated by ChatGPT. Through a meticulous exploration of these conversations, our goal is to illuminate the potential of ChatGPT to enhance software development practices. This research promises to provide valuable insights into the evolving role of ChatGPT in the world of software development.CCS CONCEPTS• Software Engineering → Software Quality; Refactoring.