International Journal of Computer Applications (0975 – 8887) Volume 80 – No.9, October 2013 41 Towards a Self-service Data Analytics Framework Mohamed M Zaghloul Computer and Systems Department, Faculty of Engineering, Mansoura University, Egypt Amr Ali-Eldin Computer and Systems Department, Faculty of Engineering, Mansoura University, Egypt Mofreh Salem Computer and Systems Department, Faculty of Engineering, Mansoura University, Egypt ABSTRACT The need for Self-service data analytics is inevitable as it supports the business in making the right decisions. In this paper, we argue that self-service analytics frameworks should be based on a process-centric approach and visualized self- service components in order to meet current business demands. Further, we enunciate the need for mainly three components: Map component, Process Flow component and a Control Model component. Furthermore, we explain the architecture of a self-service analytics framework based on these components. Some parts of the proposed framework were deployed to different sites and are discussed in detail in this paper. The obtained results showed a clear enhancement of data warehouse operation spent from the IT departments' side compared to the traditional BI architecture. General Terms Self Service Business Intelligence, Self Service Data Analytics . Keywords Extraction, Transformation, and Loading (ETL), Business Intelligence (BI), Process-centric collaboration, Self-service data analytic, Control model Operational Data Store (ODS) 1. INTRODUCTION Business Intelligence provides corporate with analytical reports that transform data into information enabling business users to take the right decisions at the right time [1], [2]. This model provides corporate with their competitive edge and allows them to generate new business opportunities. However as per Gartner reports [3], the current BI environment is not an enabler for such ambitions and is still under the required level in a way that does not cope with the ever-evolving business changes [3]. As per Gartner, Self-service BI is defined as business users being able to generate the reports they need throughout their daily work cycle without seeking the help of their IT department [3], [4]. This means the IT should be separated from the business, which is very useful as it will lead to quicker access to data and the ability to transform it into information faster; however, it is not that simple [3], [4]. It cannot be expected that business people will be given the latest BI technologies and asked to avoid working with IT; they will be frustrated as tools contain many complexities they cannot address by themselves. Business users will face problems like data quality and consistency, query performance scalability, complex data mining algorithms needed to produce more advanced analytics, the need to integrate more data sources to respond to new business requirements, and much more [3], [4]. This is actually what this paper is trying to address. The researchers are suggesting a new Self-service business analytic framework able to eliminate the difficulties business users will face, and consequently enable Self-service business analytics to become a useful approach corporate can start applying successfully. This paper organized into the following sections; Section 1 presents a short Introduction. Section 2 discusses the research challenges and describes traditional architecture challenges. Research objectives are described in section 3. The traditional BI architecture approach is presented in section 4 while the proposed Architecture is presented in section 5. Experimental work of the proposed framework is described in section 6 followed by an analysis of the results and a discussion in section 7. Section 8 presents related works and afterwards the paper is concluded in section 9 .Section 10 References .Section 11 Appendix. 2. RESEARCH CHALLENGES So what are the challenges that meet business users while using the traditional BI software and that the suggested framework will be able to eliminate. 2.1 Changing Business Demands Business Intelligence software is accused by not being able to cope with the business analytics requirements that continuously change every now and then. As business grows, the data traffic becomes higher in the corporate transactional systems, and thus new analysis requirements start to arise. The analytics requirements that BI software is actually covering compared to the real analytics needs become lesser every day. This is due to the fact that BI software is dealing with an underlying set of data marts responsible for generating these business analytics. Business users are not able to extend the existing data marts to include new data needed, and thus generate the new required analytics. To do so, they have to refer to their IT department, and here starts the latency and the cost increase. 2.2 High Operational Costs Evolving business changes and the new analytics needs they incur also lead to issues related to the ETL (the technique responsible for extracting data and transforming it into information) [4]. In order to cope with the increased volume of data, the ETL may start facing issues related to performance and accuracy.