Requirements-driven modeling for decision-making systems
Nawfal El Moukhi, Ikram El Azami, Abdelaaziz Mouloudi, Abdelali El Mounadi · 2018
A data warehouse allows to integrate heterogeneous data sources for analysis purposes. One of the keys of success of the data warehousing process is the design and definition of the warehouse model based on the requirements of decision makers. This paper presents and analyzes the set of requirements-driven methods for designing data warehouses and classifies them into two broad categories: users-driven and goals-driven methods. Then, it introduces a new method called XCube Assist which allows to exploit all advantages of the users-driven approach. Our new method is mainly based on the semantic analysis of the requirements expressed in natural language and their confrontation with data sources and the search history. At the end of the process a multidimensional model is generated in XML language.