Sdmx5 modelling a statistical domain
Download as PPTX, PDF0 likes236 views
This document discusses the steps to model a statistical domain in SDMX, including agreeing on data exchange needs, defining a concept scheme with all required concepts, coding the concepts, defining a DSD matrix relating concepts to dataflows, optimizing the matrix to reduce dimensionality, and creating the final SDMX artifacts like the concept scheme and dataflows. The key steps are to identify data exchange needs, define all required concepts, relate concepts to dataflows in a DSD matrix, and optimize the matrix to create the final SDMX components.
Sdmx5 modelling a statistical domain
- 1. SDMX 5 Modelling a statistical domain UN/FAO - Erik van Ingen Abuja, Nigeria, 13-14 May 2015
- 2. •Agree on exchange needs •Define the Concept Scheme •Code the Concept Scheme •Define a DSD Matrix •Optimise the DSD Matrix •Create SDMX Artefacts
- 3. agree on data exchange needs list Dataflows to be covered by SDMX One Dataflow per data exchange need (e.g. transmission table)
- 4. define the Concept Scheme list all Concepts needed in any Dataflow place all of the Concepts required in a single Concept Scheme
- 5. code the Concepts add default representations to Concepts drawing on existing: cross-domain Code Lists existing shared Code Lists new Code Lists based on existing classifications new Code Lists following the SDMX coding guidelines
- 6. define a DSD Matrix put the Dataflows in relation to the Concept Scheme: which Concept is used in which Dataflow?
- 7. DSD Matrix
- 8. optimise the DSD Matrix One DSD for each Dataflow with only the Dimensions needed for the respective Dataflow A single DSD covering both Dataflows, where the code reported for Dimensions not needed are fixed to a single code for use with that specific Dataflow
- 9. Option 1
- 10. option 2
- 11. optimise the DSD Matrix fill blanks in the matrix with default codes as far as possible from statistical perspective to reduce the number of DSDs, you may fix a Dimension to a single Code value for use in a Dataflow
- 12. generic example
- 13. create SDMX Artefacts Derive DSDs from the matrix 1 Concept Scheme 1 DSD for each set of dimensionality 1 Dataflow per exchange need 1 Cube Region Constraint per Dataflow
- 14. DSD matrix in context
- 17. thank you