Content and Temporal Analysis of Communications to Predict Task Cohesion in Software Development Global Teams

Abstract

This dissertation proposal describes a study of interaction-based measures and their ability to predict cohesion within global software development projects. Messages were collected from six software development projects that involved students from different countries. The similarities and quantities of such interactions will be computed and analyzed. The tested measures will be analyzed at individual and group level. Similarly, content features based on communication categories, found in virtual learning teams, will be used to improve the identification of task cohesion level. Finally, temporal interaction similarity measures will be calculated to assess its prediction capabilities in a global setting.