SDMX 4 Dummies

This page shows in a simplified way how you can interpret a csv file in SDMX terms. It shows a CSV or Ecell file with two records. The first record has the header. The second record shows the data. In the year 1990 in the FAO Major Area (FMA) 07 there is a catch of 3244 ktonne of Tuna (TUN) , caught by Italy (ITA). Descriptions of the codes for each language can be found in the codelist.
SDMX4Dummies.png

Note

Terminology used is conform SDMX 2.1. A DataStructure is called in SDMX 2.0 a KeyFamily.
A Data Stucture Definition (DSD) is complete document containing the DataStructure, a ConceptScheme and the associated CodeLists.


Why using SDMX?

Often organisations do not use any or little standards or structure for their statistical data. In the best case they use Excel or CSV files. SDMX provides standards and structures for statistical data. SDMX is the best out there. Competing and/or complementary standards are DDI and DSPL.

Data usecases where you can use SDMX for:
  • Harmonization of classifications (codelists)
  • Harmonization of structures (Data Structure Definitions)
  • Organisation of Metadata
  • Visualization
  • Dissemination
  • Collection
  • Curation

The SDMX model is a proven model for the organisation, collection and dissemination of statistical data. Implementing SDMX may seem to be a dry and academic exercise. But the benefits of implementing SDMX is that it provides a excellent model for solving generic statistical problems and it provides a standardized interface to the outside through webservices. These webservices can be used for data collection and data dissemination.

SDMX is very popular for data collection and dissemination. Recently it is also more used for data visualisation. Good example is the this nice website of the ECB. Initiatives are on the way to integrate SDMX also with the spectacular GapMinder.

DataFlow

DataFlow.png

The diagram above shows the position of DataFlow in the relation to other important SDMX artefacts. An Agency can have more DataFlow(s). Each DataFlow has a reference to 1 DataStructure. An SDMX dataset (Data) has a reference to 1 DataStructure. A DataStructure can have more CodeList(s).





Where else to learn about SDMX?

You can learn more about SDMX on the Eurostat's SDMX Wiki. Eurostat gives sometimes free courses on SDMX for National Statistical Institutes and UN organisations. Metadata Technology offers commercial courses. You can raise questions on SDMX in the SDMX users forum. Feel free to raise any question in our forum.