Software attribute-mapping visualization as a tool in monitoring and managing software development process

Abstract

This thesis aims to examine a novel approach for software visualization. In the software development process, there is a need to do "heath checks" for the development process, software, and use. Visualization is an effective solution for cross-disciplinary communication for different software development stakeholders. Meanwhile, the traditional approach can not work with the aggregated data coming from multiple sources that can be displayed on the same visualization and can not target specific types of stakeholders. A new approach for software visualization uses an attribute-mapping technique, utilizing the visual notation concept in visualization engineering. The new approach gives the visualization software flexibility for users to define their idea visualizations by configurations expressing how the data attributes are mapped the visual variables. The stakeholders can use the new approach to rapidly create many visualizations to test whether the visualization ideas are effective. The objective is to find the best ideas of how data attributes should be represented on the visualization. It is claimed that typical software engineering projects include many repeating activities and cycles, for example in scrum or agile software development, they are called "sprints" and "delivery cycles". Those repeating tasks and iterations can be shown as repeating visual patterns. Any unusual activities of the software development process, software, and use would show unusual visual patterns that help stakeholders identify and analyze. The attribute-mapping visualization is implemented in VISDOM project, in experimenting in monitoring students' learning process in a course at Tampere University. The outcome of the experiment is the creation of a configurable visualization that works with data coming from multiple sources, intends to work with different types of specific stakeholders and has flexibility for the users to rapidly define how the data attributes manipulate the visual variables of the visualization.