Modeling of census data in a multidimensional environment

Abstract

The general aim of the KOSTRA project, initiated by Statistics Norway, is to set up a data reporting chain from the norwegian municipalities to a central database at Statistics Norway. In this paper, we present an innovative data model for supporting a data analysis process consisting of two sequential data production phases using two conceptional database schemes. A first data schema must provide a sound basis for an efficient analysis reflecting a multidimensional view on data. Another schema must cover all structural information, which is essential for supporting the generation of electronic forms as well as for performing consistency checks of the gathered in-formation. The resulting modeling approach provides a seamless solution for both proposed challenges. Based on the relational model, both schemes are powerful to cover the heterogeneity of the data source, handle complex structural information, and to provide a versioning mechanism for long term analysis.