An Improved Agile Framework For Implementing Data Science Initiatives in the Government

Abstract

Implementing data mining projects in governmental organizations is emerging in the Middle East. The literature has been showing that there is a significant gap between the problems defined by the research in data mining and the problems in real world projects. The gap is to the level of semantics between the data scientists and the business users. Trying to fill this gap, we have developed an improved Agile data mining framework to fulfill the government business objectives and needs. The previous works had been claiming that handling such project is not yet mature in the region. For this an Agile implementation framework is required. We are also proposing a systematic way for identifying business problems as part of the framework. The process is Agile, so it would start from investigating the data set dimensions to identify business problems. It also allows early Business people cooperation with data scientist. We've applied the proposed framework in one of the Middle East government organizations. The business team and the data scientists have been showing their satisfaction regarding the results of using the proposed framework. The proposed framework have helped both business and data scientist to implement their first initiative in data mining. The proposed framework also helped in efficiently mapping the project with the core business objectives and problems using real world dataset.