[Back]

M. Auer, B. Graser, S. Biffl:
"An Approach to Visualizing Empirical Software Project Portfolio Data Using Multidimensional Scaling";
in: "Proc. of IEEE International Conference on Information Reuse and Integration (IRI 2003)", IEEE Computer Society Press, 2003, ISBN: 0-7803-8242-0, 504 - 512.

Software project portfolio managers and process engineers have to manage increasingly large software project portfolios rather than single projects. Typical portfolio decisions like resource allocation, effort estimation and risk valuation are to be based on complex, high-dimensional software project portfolio data.
Handling and understanding this vast amount of data is difficult. Classical analysis approaches - like ABC-analysis for project prioritization, or interactive data mining tools - mostly analyze specific and separate subsets of the information for a given purpose. This paper (i) applies multidimensional scaling methods for visualizing high-dimensional portfolio information, involving data from many different sources, and describes promising applications of the approach in this domain, (ii) proposes a simple method to prepare raw software metric data for MDS analysis and processing with interactive tools, and (iii) provides a feasibility study based on a small set of open-source software projects to demonstrate the usability of the proposed approach for software project portfolio decision support.
The approach enhances explorative analysis of historic portfolio information, allows for interactive data analysis with intelligent tools, and thus directly supports portfolio assessment and prediction decisions.

http://aleph.ub.tuwien.ac.at/F?base=tuw01&func=find-c&ccl_term=AC04404452