Discovering patterns of correlation and similarities in software project data with the Circos visualization tool

TL;DR

This paper introduces the use of the Circos visualization tool, borrowed from bioinformatics, to identify patterns of correlation and similarity in software project data, aiding in cost estimation model development.

Contribution

It proposes an innovative visualization approach for software engineering data analysis, enhancing pattern discovery and preliminary data exploration.

Findings

Circos effectively visualizes correlations in software project data

The tool reveals project similarities and variable relationships

Facilitates better data understanding for cost estimation models

Abstract

Software cost estimation based on multivariate data from completed projects requires the building of efficient models. These models essentially describe relations in the data, either on the basis of correlations between variables or of similarities between the projects. The continuous growth of the amount of data gathered and the need to perform preliminary analysis in order to discover patterns able to drive the building of reasonable models, leads the researchers towards intelligent and time-saving tools which can effectively describe data and their relationships. The goal of this paper is to suggest an innovative visualization tool, widely used in bioinformatics, which represents relations in data in an aesthetic and intelligent way. In order to illustrate the capabilities of the tool, we use a well known dataset from software engineering projects.