Database system support for multidimensional data analysis in environmental epidemiology

The project CARLOS (Cancer Registry Lower-Saxony) aims at novel techniques for analysing cancer clusters using advanced database technology to support multidimensional analysis of spatio-temporal related health data. This requires very advanced support capabilities which can not be provided by existing technology. Therefore, new approaches in the fields of spatio-temporal, multidimensional and statistical database systems have to be considered to develop an appropriate support architecture. We present our experiences and ideas gained in implementing a prototype. Based on these facts, we describe our approach to dynamically create statistical objects over predicate based aggregates in the course of epidemiological data analysis