Towards Intelligent Analysis of Commplex Networks in Spatial Data Warehouses

This paper proposes a knowledge-based model to support spatial and temporal analysis of complex networks in data warehouses. The proposed model, inspired from geography, represents the dynamics of spatial elements in the network. Ontologies, describing element classes from a domain, space partitions, and time periods, respectively, help to define dimensions of data marts customized for specific analysis needs. This model has been implemented in a prototypical framework, with a GUI supporting spatial OLAP and visualization of the varying state of network portions in graphs and maps. It supports the investigation of spatial and temporal patterns and tendencies, as has been observed in an electricity distribution case study.