Abstract:
Mass software development requires the collaboration of many people with miscellaneous
technical backgrounds as it is a very complicated problem-solving process. The technical
perspectives play a vital role in the success of a software project, but it also depends heavily
upon how efficient and effective those people operate as a team. The key success factors of
a software project are recognized as proper team and time management. But most of the
time software projects fail to meet their initial timelines and to deliver the product features
to meet the original objectives. The conventional models which are currently in use for team
composition and time estimation in software projects are mostly based on data gathered in
the design phase of a project. The involvement of project managers in manually feeding the
data into the models compromises the accuracy of the output of the model. Also, these
models were introduced many years back and the parameters used in the models are
outdated due to the constantly changing software development methodologies. The
existing algorithmic approaches used to solve the team composition and time estimation
problem in software projects are mainly based on team size. Synergy, the proposed data
analyzing statistical model will facilitate team composition and time estimation in software
projects by mining contribution history of developers and projects on GitHub which is a
vastly growing source code management platform for software projects. The model is based
on the parameters to assess the prior experience of programming languages and source
code contribution volume of the developers and the contribution history of completed
projects. This dissertation will provide an analysis and evaluation of the capabilities of the
proposed statistical model in optimizing the accuracy and the efficiency of team
composition and time estimation process for software development projects.
