Intuitive: modelo conceitual para workflows de ETL

Abstract

The information domain is seen as a competitive differential in the most varied business areas, such as health, agribusiness, telecommunications, logistics, and government agencies. The correct and updated information is a valuable subsidy for corporative strategic decisions. Additionally, nowadays, huge volumes of data are generated at high speed and in various formats. In this context, research has been made to propose new models, architectures, processes, and algorithms that can contribute to transforming data into useful information for strategic decision making. In this scenario, a data warehousing environment plays a key role. The environment contains the data warehouse (DW), a huge repository with data that serves as a basis for responding to OLAP (Online Analytical Processing) queries. In a data warehousing environment, the ETL process is used to extract raw data from different data sources and to transform, clean, and integrate that data, loading to the DW. The ETL process is used for first data loading and, also for refreshing the data in the DW. This master's research investigated the best practices in conceptual modeling for ETL workflows and, as a result, proposes a new model, called “Intuitive”. The Intuitive Model adds simplicity, agility, clarity, and consistency to the modeling stage and can contribute to the improvement of construction and maintenance of ETL workflows. Theoretical analysis activities and practical experiments were performed with the users’ participation in order to validate the Intuitive Model. Such steps allowed us to evaluate that the elements of the Intuitive Model are sufficient to represent clearly several regular ETL scenarios showing advantages in comparison with the main related work in the state of the art.