CoMoTo: the collaboration modeling toolkit

We are excited to introduce CoMoTo -- the Collaboration Modeling Toolkit -- a new, web-based application that expands and enhances well-known software similarity detection systems. CoMoTo is an end-to-end data management, analysis, and visualization system whose purpose is to assist instructors of courses requiring programming exercises to monitor and investigate the extent of student collaboration, both allowed and illicit. We describe CoMoTo's interface, which was designed to facilitate scrutiny of collaboration data projected along student, course, assignment, etc. attributes, and to allow for interactive visualization of pairwise similarity measures via a dynamic graph. We also elaborate on the details of CoMoTo's implementation. Finally, we briefly discuss two use cases that foreshadow CoMoTo's broad utility in student code analysis, not only for plagiarism detection, but also for investigating early student coding styles, and for evaluating software similarity detection systems, themselves.