Bridging context and data warehouses through ontologies

Nowadays, we are assisting to three continuously demands from companies: (i) developing analytical applications around Data Warehouse systems (DW) from numerous data sources, (ii) the explicitation the semantic of these sources to reduce heterogeneities and (iii) contextualization of sources. By examining the literature, we identify the existence of several efforts attempting to offer solutions merging these three issues. The merging has been performed partially. To be more concrete, we have identified that the two first demands have been merged. Similarly, the second and the third ones gave raise to contextual ontologies. Unfortunately, all three are not well merged. This paper proposes a comprehensive methodology to design multi-contextual semantic DWs. Our approach consists first in merging context and ontologies and then with DWs. Firstly, a connection between ontologies and context model is built at meta model level. Secondly, a formalization of multi-contextual semantic data warehouse is given, followed by a deep description of the most important steps of the data warehouse design. Finally, a case tool and experiments are conducted using a contextualized hospital ontology to show the effectiveness of our approach.