Influence of Communication Among Shared Developers on the Productivity of Open Source Software Projects

TL;DR

This paper investigates how communication among shared developers influences open source project productivity by analyzing GitHub data, focusing on issue resolution times and feature importance using polynomial regression.

Contribution

It introduces a data-driven approach to identify key features affecting productivity in open source projects, utilizing API data and advanced preprocessing techniques.

Findings

Communication features impact productivity

Polynomial regression effectively models issue resolution times

Preprocessing improves model accuracy

Abstract

Many software developers rely on open source software for developing their applications and writing their source codes. Measuring an independent project's overall productivity is still an open problem for many technology companies. In this project, we address to bridge the gap of analyzing which are the most important features for prediction of a productivity based system. We have chosen to collect data from GitHub via their application programming interfaces (API) and analyze the data we gathered to understand the relation between the average time to close an issue and the features that we collected. Since most of the data we gathered were not Gaussian, we had to preprocess the data using outlier detection and applying transformations before statistical modeling. The best model we observed was polynomial regression with degree 5. Overall, we noticed that there are many aspects of software development that make developers increase their productivity.