Design, Development and Deployment of SETAP WWW Site and ML Training Database
Nidhi MiglaniOral Defence Date:
Wednesday, May 25, 2016 - 11:00Location:
Profs. Dragutin Petkovic, James Wong and Lecturer, Marc Sosnick
Goals of SETAP project are to explore machine learning (ML) approach to predicting and assessing learning of teamwork in Software Engineering (SE) education. The most complex and time consuming component of SETAP is the collection of teamwork behavior data from students working on class projects, which was done over 5 semesters in a joint SE class between SFSU, Fulda University in Germany and Florida Atlantic University (FAU) Florida. This data constitutes ML training database and consists of teamwork data collected from 74 student teams with over 380 students in the period from Fall 2012 till Fall 2015. This training data is used to train ML to attempt to predict the student teamwork learning from their behavior data. The data had to be well organized and curated in order to be effectively used for ML research. One of the key deliverables of SETAP is a Web site for dissemination of the project information and collected ML training data for use by others with the detailed documentation about the collected data. Our report covers our work related to the above two deliverables, namely: a) design, development and deployment of the SETAP Web site, b) organization, curation and testing of the data; and c) development of the detailed SETAP data documentation. SETAP WWW site is now operational and the data documentation has been posted (ML training data itself will be posted after our research publication in October 2016). The project has been supported by NSF TUES grant # 1140172.