sdmx3mdt
commented
3 years ago The SDMX Technical Working Group (TWG) reviewed the comment at the 27 July 2021 Public Consultation Final Review Meeting.
Decision: The suggested editorial enhancements to the SDMX technical materials will be taken into consideration as part of the planned forthcoming project to modernise the Standard's published documentation.
From Narodowy Bank Polski Received via twg@sdmx.org 13 July 2021
Dear SDMX leaders,
I have the following suggestions to materials about SDMX-3.0:
I. Suggestion: It is important to give the reader of the manual at the beginning clear picture about purpose of SDMX initiative. Therefore I suggest to put in introduction following sentence:
The purpose of SDMX (Statistical Data and Metadata Exchange) initiative is to develop: Transparent Notation for Statistical Information Production.
The phrase “Transparent Notation for Statistical Information Production” is better understandable for business area than phrase “Statistical Data and Metadata Exchange” (which is more technical). This definition is closer to tasks defined for Statistical Departments in different institutions.
II. Suggestion: It is important to put emphasis on transparency of SDMX notation and clearly define what does it mean transparent notation.
Intuitive definition of transparent notation is:
Technical requirement equivalent to this intuitive definition are following:
Transparent notations are very useful for statistical analysis and conclusions.
III. Suggestion:
It is important to list most important use-cases for SDMX notation: 1) Statistical Data Storage 2) Statistical Data Navigation 3) Statistical Data Description 4) Statistical Data Tabular Presentation 5) Statistical Data Exchange 6) Statistical Data Selection 7) Statistical Data Transformation 8) Statistical Data Validation 9) Statistical data Aggregation 10) Documentation of Statistical Data Transformation (drill down from calculated value by transformation formula to source observations) 11) Table Definition Queries in order to get table with economic indicators.
Therefor I suggest to include in documentation list of important use-cases.
IV. Suggestion: It would be valuable to define list of principles of SDMX 3.0. I think that number of principles should be equal to the number of use-cases. I have found 8 principles but I think there should be 3 principle more. The suggested 8 principles are below:
1) Record is primary data structure for SDMX 3.0. (In previous versions of SDMX the Time Series was primary data structure for SDMX) The set of records built according to the same rules is called Recordset. (in current and previous versions Recordset has similar definition).
2) Observation is identified by 3 parameters:
3) Record Identifier is constructed according to the pattern: Code of Recordset Identifier and sequence of code values for dimensions separated by “.”.
4) SDMX Data Exchange is performed by files containing set of records, each record contains Record Identifier, Date of observation, Sequence of measures and relevant values for each measure.
5) SDMX Observation Variable is data structure designed for SDMX data selection. Observation variable is a structure which contains:
6) Transformation formula is data structure designed for SDMX data transformation. SDMX data transformation formula is formula built on observation variables.
7) SDMX Tabular Template is a table containing cells assigned to Observation Variables. The Tabular Data Presentation is achieved by substitution of parameters by values of parameters.
8) SDMX Data Validation is logical condition for calculated value of transformation formula, it is true when validation is fulfilled and false when not fulfilled.
Kind Regards Jan Kaczanowski