Last modified by Harold Kroeze on Fri, 10/02/2017 - 14:17
Statistical Data Warehousing planning allows unknown inputs and outputs, because dealing with a new type of input or output is just a matter of adding a new conversion process. There is both metadata linked directly to the data, and metadata for the process management.
The administration of the Statistical Data Warehouse chooses how to deal with the registers of the statistical units which can be integrated in the data warehouse, replicated periodically or even considered as just another input at times. The way in which the data warehouse production affects the statistical unit register is what is relevant, and not where it is physically located.
At the statistical offices data from several different sources is gathered and many compatibility problems have to be solved in order to harmonize the data and deal with the conflicts in the statistical unit structure itself.
This is one of the major challenges but also one of the main benefits of S-DWH . The process by which the data is integrated to the common unit structure during the processing phase, involves more compromises and articulation than technical difficulties. It’s very hard and time consuming but data integration is done only once and the integrated data can be then extracted by several users, and extensively reused. In this way data validation in the processing phase increased the data quality and all the outputs coherence.