In response to demands of the software engineering industry, this project is developing powerful machine learning-based methods to understand, assess and predict student learning of software engineering teamwork across globally-distributed teams. The project includes the following activities. Objective and quantitative teamwork data are being collected on student activities in ongoing, jointly-taught undergraduate computer science classes at three geographically-distant institutions (San Francisco State, Florida Atlantic, and Fulda Universities). Novel machine learning tools are being applied to these data to discover models, rules and metrics that define student success in acquiring teamwork skills and that facilitate early intervention for teams at risk by identifying predictors of teamwork outcomes. With input from external evaluators, the methods and tools developed in this project are being refined and disseminated to educators for early adoption.
The project advances the field of software engineering with new tools for assessing student software engineering teamwork skills. This project is the first to apply novel machine learning techniques to assess and predict student learning of these skills. In the era of global collaboration, the project is also assessing the impact of developing teamwork skills within globally-distributed teams. The participating US universities serve highly-diverse student populations; thus the project is preparing large numbers of students underrepresented in STEM fields to enter the software engineering profession and to successfully meet the challenges of communicating effectively in globally-distributed teams. As a result, the project is contributing both to diversifying and to maintaining the competitiveness of the US software engineering workforce.
A training workshop for software engineering educators from across the California State University system and from a San Francisco community college is part of the project.
