There are four great restrictions in project management: cost, quality, schedule, and scope. In software project management, there is another variable: creativity. Therefore, to succeed in the entire life cycle of a project, one needs the ability to predict uncertainty, to plan activities to include time restrictions and scarce resources, and, moreover, to manage all these tasks to ensure success.
In this paper, Santos and Belo use data mining tools to create a model that could be deployed in any project management process, assisting project managers in planning and monitoring the state of a project (or program) under their supervision.
The study is based on the cross industry standard process for data mining (CRISP-DM) methodology. They use CRISP-DM to analyze relevant data extracted from a database of 24,000 projects. The selected projects were from more than 30 different countries.
Initially, they collected data from 5,000 projects and arranged them by type of industry, indicators of complexity, experience and project context, project size, resources, and so on. They used function point analysis and lines of code to measure the size of the projects.
This paper discusses a data mining application that addresses the effort of estimate on software engineering projects. The authors describe the methods they used to explore potential correlations and influences from relevant parameters such as experience, complexity, organizational maturity, and project innovation. A relational database with more than 200 tables enabled the authors to understand the business of managing software engineering projects fairly completely.
The authors contend that the model can be implemented in various software engineering projects as an alternative tool to the techniques and methods commonly used. The evaluation shows that implementing this model for risk analysis at the planning stage, as well during project execution, will enable the use of “what if” scenarios and testing. This should enable managers to measure and validate several alternatives to improve the success of projects.