Intégrer les LOD dans un cube de données : Transformons une action technique en valeur organisationnelle
In EDA 2019, vol. RNTI-B-15, pp.61-76
With the proliferation of Linked Open Data (LOD) sources, organizations have quickly seen the opportunity to extend their internal data cubes with such external data. Although LOD sources have a real added value, they have brought their own technical difficulties related to the management of this new source. However, the existing literature focus on managing these technical difficulties, whereas this decision is primarily 'organizational'. This omission is explained by the fact that the integration of LODs is studied according to a source-driven vision, whereas a requirements-driven vision would be more adequate to study effectively the relevance of the integration of LOD sources. In this paper, we propose an organizational (goaloriented) vision in which Key Performance Indicators (KPIs) are the central element linking data sources to organizational objectives. We propose a goal-oriented approach to estimate the organizational value (via KPI values) of the LOD resources in the cube. The approach is based on a metamodel aligning the three models: the data model, the goal model and the KPI model. Experiments are conducted to validate our approach.