GIST – Gaining Actionable Insights from Software Testing

GIST will provide companies with tools to support critical operational and strategic decisions. Similarly, GIST will benefit research as the validated tools can be generalized and adapted to other fields. The layered modular architecture of GIST supports reuse, improves data availability and speeds up further advances in software analysis by enabling independent contributions at different levels of the GIST architecture.

Purpose of the Project

In software development, data and visual analytics have been proposed to interpret large amounts of multidimensional data.
However, in software testing, the approach is not well researched. Moreover, the cost of existing approaches is too high for practical use due to the following limitations: (a) tight coupling between the analytics and the underlying data collection mechanisms and (b) the need to interact with and integrate different data sources to collect all the necessary data.

In GIST, we develop and evaluate tools to address these limitations. Using design science, we identify and consolidate the information needs of key stakeholders and develop analytical solutions to meet these needs. We then develop an information model to decouple the analytical solutions from the data sources and develop an efficient and reusable data collection mechanism.

Implementation

GIST will provide companies with tools to support critical operational and strategic decisions. Likewise, GIST will benefit research as the validated tools can be generalized and adapted to other fields. The layered modular architecture of GIST supports reuse, improves data availability and speeds up further advances in software analytics by enabling independent contributions at different levels of the GIST architecture.

Our team is an ideal combination to successfully implement GIST. Our corporate partners from Axis, Ericsson and TestScouts have a long experience in software testing. Their domain knowledge and familiarity with research collaborations benefit the project. On the academic side, we have extensive experience with industrial collaborations, software analysis, empirical research and software testing. We also have relevant experience in data warehousing and software development.
While GIST focuses on software testing, the solutions can also be generalized to other areas of knowledge in software engineering. It will thus create synergies with ongoing local and national research projects. The experience gained from the project will also be continuously integrated into our teaching and lead to theses on related topics.