Characterizing relationships for system dynamics models supported by exploratory data analysis : A conceptualizing approach about the meeting diversity in student software projects

Abstract

Estimating dynamic components in projects involves understanding human factors which are substantial in software development. Communication and collaboration in teams consist of social-driven characteristics with influences on the continuous delivery of software. Efficiently estimated meetings become increasingly important due to budget calculations and shortened release cycles. Experiences of project managers combined with retrospectives on historical data records support a better understanding of team dynamics. But interpreting complex effects is not always trivial, in particular without further analyzes. In several studies, information relationships are investigated through linear correlation measures. Additional analyses for higher correlations are often neglected due to the advanced functional characterization. This leads to statistical gaps with significances for explored data relationships and their functional interpretation. In this paper, we present a systematic identification and visualization of team communication effects and diversities for field study records of 34 student software projects. We combine methodologies from system dynamics with exploratory data analysis to extract and emphasize significant effects. These insights help to sensitize for advanced investigations about the statistical measures of correlation and to interpret sophisticated structures. Furthermore, it reinforces potentials for a team's communication performances and enables an enhanced understanding about how student teams meet and communicate.