Konferenzbeitrag
MUSTANG — A Framework for Statistical Analyses of Multidimensional Data in Public Health
Vorschaubild nicht verfügbar
Volltext URI
Dokumententyp
Text/Conference Paper
Zusatzinformation
Datum
2003
Autor:innen
Zeitschriftentitel
ISSN der Zeitschrift
Bandtitel
Verlag
Metropolis
Zusammenfassung
In environmental science and public health, the need for sophisticated analyses on multidimensional data with regard to geographical properties is an established fact. Unlike data analysts in business contexts, scientists not only use typical OLAP-operations on multidimensional data, but also domain-specific complex statistical procedures like spatial cluster analyses. In addition, expectations on design style and complexity of user interfaces vary from community to community.
Our framework MUSTANG meets these requirements and furthermore standardises the development of domain-specific data analysis tools. This multi-tier framework is based on a well-defined component for processing multidimensional data. A second component computes statistical procedures on multidimensional data and enables software developers to plug in the statistical procedures required in the scenario the framework is used in. A special API on top of the framework offers a process oriented object model with objects like investigations, tables, and maps. Application programmers may use this interface to develop platformand domain-specific user interfaces.