XML Source Preparation for Building Data Warehouses

Faced with the high economic competition, today’s enterprises are forced to rely on decision support systems to assist them in the analysis of large data volumes. Traditionally, the analyzed data are mainly issued from the enterprise’s operational information system. However, due to the international nature of the competition, enterprises are increasingly pressed to explore other, external data sources, mainly issued from the web. However, despite the convergence of XML as a standard data format on the web, a main difficulty in exploiting these data is the lack of data warehouse design approaches for XML data sources. The few proposed approaches suppose that the designer must first manually identify the XML element that represents the analyzed fact; such a step requires a high expertise and domain knowledge. The main contribution of this paper is to automate this step.