Software portfolios of today's companies are comprised of a variety of heterogeneous, modular, and often polyglot software solutions. Ensuring high-quality standards across these entire portfolios raises additional challenges for quality engineers and requires new strategies and visualization approaches to support software quality management decisions. In this paper, we study the information needs of software quality engineers that drive these decisions on portfolio-level and propose Portfoliotrix1 - an expert visualization prototype for a portfolio-wide analysis and comparison of software quality metric trends. For this, we first introduce our four-phase fully automated data-mining process for gathering and aggregating quality metric data from the underlying software repositories. We then present our application frontend that implements different visualization and filtering concepts to satisfy eight specific information needs derived from our prior qualitative semi-structured expert interviews. A final scenario-based expert evaluation confirms the practical relevance of our prototype. Experts took less time to complete quality-related management tasks with our tool than they had estimated when relying only upon existing toolsets. Moreover, our visualization system reached the System Usability Scale (SUS) score of "good" with 76.7 points.

1 Artifact is available at https://doi.org/10.5281/zenodo.5475463

Software portfolios of today's companies are comprised of a variety of heterogeneous, modular, and often polyglot software solutions. Ensuring high-quality standards across these entire portfolios raises additional challenges for quality engineers and requires new strategies and visualization approaches to support software quality management decisions. In this paper, we study the information needs of software quality engineers that drive these decisions on portfolio-level and propose Portfoliotrix^1^ - an expert visualization prototype for a portfolio-wide analysis and comparison of software quality metric trends. For this, we first introduce our four-phase fully automated data-mining process for gathering and aggregating quality metric data from the underlying software repositories. We then present our application frontend that implements different visualization and filtering concepts to satisfy eight specific information needs derived from our prior qualitative semi-structured expert interviews. A final scenario-based expert evaluation confirms the practical relevance of our prototype. Experts took less time to complete quality-related management tasks with our tool than they had estimated when relying only upon existing toolsets. Moreover, our visualization system reached the System Usability Scale (SUS) score of "good" with 76.7 points.

^1^ Artifact is available at [https://doi.org/10.5281/zenodo.5475463](https://doi.org/10.5281/zenodo.5475463)

Visualizing Metric Trends for Software Portfolio Quality Management

The ninth IEEE Working Conference on Software Visualization (VISSOFT 2021) builds upon the success of the previous eight editions of VISSOFT, which in turn followed after six editions of the IEEE International Workshop on Visualizing Software for Understanding and Analysis (VISSOFT) and five editions of the ACM Sympo0sium on Software Visualization (SOFTVIS). In 2021, VISSOFT will take place and be co-located with ICSME.