A model driven process for spatial data sources and spatial data warehouses reconcilation

Abstract

Since the data warehouse integrates the information provided by data sources, it is crucial to reconcile these sources with the information requirements of decision makers. It is specially true when novel types of data and metadata are stored in the data sources, e.g. spatial issues. In this way, spatial requirements have to be conformed with the available spatial metadata in order to obtain a data warehouse that, at the same time, satisfies decision maker spatial needs and do not attempt against the available metadata stored in the data sources. Therefore, in this paper, we have based on multidimensional forms and some spatial and geometric considerations to define a set of Query/View/Transformation (QVT) relations to formally define a set of rules that help designers in this tedious and prone-to-fail task. The novelty of our approach is to consider an hybrid viewpoint to develop spatial data warehouses (SDW), i.e., we firstly obtain the conceptual schema of the SDW from user requirements and then we verify its correctness against spatial data sources by using automatic transformations. Finally, the designer could take decisions to overcome the absence or incompactibility of certain spatial data by using our Eclipse CASE tool. © 2010 Springer-Verlag Berlin Heidelberg.