TravisTorrent is a data set synthesized from GitHub and Travis CI that provides access to hundred of thousands of builds from a thousand of selected projects. Mining it researchers can evaluate hypothesis regarding the use of CI, testing patterns and software quality.
Visualization is suitable for the exploration of relationships among multivariate data, such as TravisTorrent. In this seminar project we will define hypothesis that we will assess through lightweight visualization. Using the Pharo language and the Roassal visualization engine, we will define an agile pipeline for fast test of hypothesis. The outcome of the project is expected to contribute to MSR 2017 Challenge.
Contact Leonel Merino