The Metrics to Evaluate the Health Status of OSS Projects Based on Factor Analysis

Abstract

As open-source software (OSS) development is becoming a trend, an increasing number of businesses and developers are joining OSS projects. For project managers, developers and users, understanding the current health status of a project is very important to manage a development process, select the open-source projects to development or to adopt the software packages developed by projects. Therefore, an efficient approach to evaluate the health status of the open-source project is needed. Unfortunately, although many approaches including metrics have been proposed, they are designed in arbitrary ways. In this paper, a math ematical tool, i.e., factor analysis, is used to build a health evaluation model for OSS projects. As far as we know, this is the first time that factor analysis has been applied to evaluate OSS projects. This model is based on GitHub data and uses the basic indexes that are closely related to the health status of the projects as the input. Then, six new synthetic metrics, namely community activity, project popularity, development activity, completeness, responsiveness and persistence are obtained through factor analysis, which can be used to calculate the overall health score of a project. Moreover, in order to verify the effectiveness of this model, it is applied to some real projects and the results show that the overall scores achieved by this model can reflect the health status of the projects.