DIME/ITDG Plenary February 2015/09/EN
Joint DIME/ITDG Meeting
24 February 2015
Item 9 of the agenda
The ESS reference metadata standards
5
The ESS reference metadata standards
1. Recommendation for action by the Joint meeting
The Joint meeting of DIME and ITDG is invited:
§ To approve the revised versions of the ESS reference metadata reporting standards (ESMS 2.0 and ESQRS 2.0) as part of SIMS and the timetable for the implementation;
§ To support the plan for further standardisation of European and national reference metadata relating to European statistics;
§ To give feed-back on the work on improving the technical interoperability between the national reference metadata systems and the European system.
2. Background and brief history
The joint DIME and ITDG meeting of 26-27 March 2014 discussed the ESS reference metadata standardisation work and in particular the need to define an implementing strategy for the Single Integrated Metadata Structure (SIMS) as required by the ESSC at its meeting in November 2013. The joint DIME and ITDG meeting of March 2014 took note and generally supported the standardisation, including the principle of a common reporting on metadata and quality.
The DIME and ITDG of March 2014 and the ESSC of November 2013 also gave the feed-back to minimise the domain specific metadata structures and the more technical work to be carried forward in the Task Force on implementation of metadata standards[1] and in the Working Group on Metadata.
This document presents the work of the Task Force on metadata standards implementation as carried out in 2014 and the conclusions of the Metadata Working Group from its meeting in July 2014.
3. Policy context
3.1 Revision of SIMS and alignment of ESMS and ESQRS
The Euro SDMX Metadata Structure (ESMS) and the ESS Standard for Quality Report Structure (ESQRS)[2] were originally developed for different purposes and are as such are not fully consistent with each other. Since ESMS as well as ESQRS are now the ESS reference metadata reports structures derived from SIMS[3], the consistency between these report structures needs to be improved/ ensured. This section reports on the proposal for an amended ESQRS and the implied consequences evolving for ESMS and SIMS.
Revision of ESQRS
Concerning the order of concepts, a quality report should start with the presentation of the data (Statistical presentation with all its sub-concepts) and the processing (Statistical processing and its sub-concepts). Afterwards the Quality management issues are introduced, followed by the output quality dimensions according to the Code of Practice (Relevance, Accuracy and Reliability, Timeliness and Punctuality, Coherence and Comparability, Accessibility and Clarity). Information on Cost and Burden, Confidentiality and an item for additional Comments close the report.
Apart from the order, with regard to the concepts and sub-concepts the following modifications for ESQRS are suggested:
§ The former concept “Introduction” will be substituted by “Statistical presentation” and its sub-concepts;
§ The concept “Quality management” and its sub-concept “Quality assurance” will be added.
§ “Data revision – policy”, “Data revision – practice” and “Data revision – average size” are included as sub-concepts of “Accuracy and reliability” because they refer directly to reliability. The 1-digit concept “Data Revision”, was only a heading, is not needed as it can lead the reader to think that Data revision is something distinct from "Accuracy and reliability". It is therefore removed.
§ The concepts “Coherence” and “Comparability” should be merged in the same heading, as they appear in the Code of Practice. The sub-concept “Comparability-domain” should be removed.
Revision of ESMS
The item related to the quality concept "Accessibility and Clarity" has a slight impact on ESMS. In order to harmonize both, the report structures ESMS and ESQRS, the concept “Accessibility and Clarity” should be used by merging the current concepts "Dissemination format” and “Accessibility of documentation". Furthermore, the concepts “Comparability” and “Coherence” should be merged to one concept “Coherence and Comparability”. Due to the above mentioned modifications the concepts in the ESMS should also be reordered.
Revision of SIMS
The SIMS as a dynamic inventory and conceptual framework for all ESS quality and reference metadata concepts, with a unique definition and clear reporting guidelines should then be updated. The concepts “Introduction”, “Dissemination format” and “Accessibility of documentation” should be removed accordingly. The concepts “Coherence” and “Comparability” should be merged to one concept “Coherence and Comparability” whereas the sub-concept “Comparability – domain” should be removed.
The revised versions of ESMS (ESMS 2.0) and ESQRS (ESQRS 2.0) with the changes highlighted as worked out by the Task Force on implementation of metadata standards, and approved by the Metadata Working Group on 1 July 2014, can be found in annex 1.
If the revised standards are approved by DIME and ITDG, the preparatory work for the changeover will start. The impact of the changeover to the version 2.0 of the standards is for the users of the metadata minor. The impact (and cost) will mainly be for adapting the existing European and national metadata systems to the new versions of the standards. The impact on the metadata systems has been discussed in the Task Force on implementation of metadata standards. The Task Force considered it feasible to adapt the metadata systems to the new standards by allocating sufficient time for the changeover (i.e. six months was considered realistic for this work).
3.2 Strategy for implementing the ESMS/ESQRS standards in statistical domains
In the document presented to the DIME/ITDG meeting of March 2014, two main recommendations were made related to the coverage of national ESMS flows and to the coverage of national ESQRS flows for the statistical processes related to European statistics. The first recommendation stated that all statistical processes producing, exchanging and disseminating European statistics should be accompanied by general reference metadata – including basic quality information according to the ESS quality criteria –using the ESMS. The second recommendation stated that those statistical processes producing, exchanging and disseminating European statistics that require more detailed reference metadata on the different quality concepts should use the ESQRS.
Based on these recommendations, Eurostat and the Task Force on metadata standards implementation have worked on the further European and national reference metadata standardisation related to the European statistics:
§ The criteria to decide if national reference metadata should be implemented either according to ESMS or ESQRS are for instance legal bases and requirements, the availability of already collected metadata on quality, the characteristics of the statistics (derived or primary statistics) or the frequency of data collections;
§ The discussion on further implementation of metadata standards should deepen. The mere notion that “ESMS or ESQRS is not relevant” for a certain statistical process (for instance due to specific metadata reporting requirements defined in domain specific legislations) cannot be sufficiently justified and is not a valid statement. ESMS or ESQRS must be implemented for all statistical processes but more investments are needed in the implementation of the standards in order to improve communication about the standards and support the accommodation of the needs of the statistical domains, preferably within existing structures;
§ The work in progress to establish a reporting structure for the ESS Inventory of Sources and Methods applicable to the different domains covered by the MIP[4] should be consistent with SIMS. Hence, also for the concerning MIP related statistical processes, ESMS or ESMS together with ESQRS must be implemented. The work will continue also to further explore and exploit the full potential of the existing standards in order to support a later decision for or against the implementation of a new standard;
§ The work of the Committee on Monetary Financial Balance of Payments (CMFB) technical group to develop a short generic template for national reports on quality for the statistics underlying the MIP should, from an ESS and Eurostat metadata perspective, be implemented through the use of ESMS and/or ESQRS in order to prevent a proliferation of standard report structures, bearing in mind the stance taken that domain specific standards should be kept to a minimum.
Coverage of statistical process
Eurostat has identified 104 statistical production processes[5]. For 80 of these, there are national data collections/ data transmissions and thus also the need for national reference metadata. The table below summaries the current state and the to-be-state.
Until now, Eurostat and countries have standardised national metadata covering 37 processes out of the total of 80 processes. It should be noted that often several ESMS reference metadata flows are opened per statistical process. For instance, 23 metadata flows (one flow per statistical indicator) have been opened for the statistical process “Short-Term Business Statistics”. Most of the time, one ESQRS is used per process.
Regarding the further standardisation, the Eurostat reference metadata team has asked the production units in Eurostat when they would consider to standardise reference metadata in their domains. Based on the provisional information provided (it depends on many factors such as the periodicity of the data collection and also discussions and agreements in the domain specific working groups), the further implementation in 2015-2018 is outlined in the table below.
Table. Current and estimated implementation of national metadata standards for the statistical processes related to European statistics
2012-2014 / 2015-2016 / 2017-2018 / TotalESMS / 14 processes (43 flows) / 16 processes / 11 processes / 41 processes
ESQRS / 23 processes (28 flows) / 9 processes / 7 processes / 39 processes
Total: / 37 processes (71 flows) / 25 processes / 18 processes / 80 processes
The upcoming implementations of the metadata standards for 2015 and 2016 will cover domains such as Time Use Surveys, LFS ad-hoc module, Structure of Earnings Survey (every four years, SES 2014 implementation in 2016), Migration statistics (2015), Household Budget Surveys (every five years, implementation HBS 2015 by 2016/17), Structural Business Statistics (2015) and Air Emission Accounts (2016). The implementations also depend on the decision to use the ESMS for meeting the needs of the generic template for national reports on quality for statistics underlying the MIP- developed by the CMFB technical group.
The details of the implementations are monitored by the domain specific Working Groups and by the Metadata Working Group.
The DIME/ITDG is asked to support a full standardisation of the ESS reference metadata by 2018. It would imply a strong commitment by both Eurostat and countries for meeting this target.
3.3 Technical interoperability between ESS and national metadata systems
Several NSIs have already adapted or are in the progress of adapting their national metadata systems to accommodate the ESS metadata standards. In order to pave the way for a more automatic exchange of metadata from the national metadata systems to the European metadata system, the Task Force on metadata standards implementation and in particular Statistics Denmark and ISTAT have worked on the technical architecture and interoperability for the exchange of reference metadata between the national and the Eurostat metadata systems. The main purpose of the work is to create a common understanding and agreement among stakeholders on the exchange of reference metadata and its components in terms of responsibility, functionality, quality and terminology.
The main document on this can be found in annex 2. The document firstly presents stakeholders, followed by an overview of the technical architecture and the connected requirements at different levels of interoperability.
4. Consequences for the NSIs
The NSIs are concerned by the actions proposed above. The increasing use of the ESS reference metadata standards in the ESS will require more reference metadata files produced and exchanged within the ESS. The adaptations of the ESS reference metadata standards (ESMS and ESQRS) will generate some additional costs also at the national level. The technical interoperability solutions need to be implemented further at the national level.
5. Outstanding problems
The outstanding problems of further harmonisation and integration of the ESS reference metadata are described in the chapters above. Decisions are needed on the proposals above in order to progress with further ESS standardisation.
6. Risk assessment
Main risks linked to the actions proposed above are:
§ A slow-down in the standardisation and business process integration for reference metadata within the ESS;
§ Stable or even increasing costs for the ESS reference metadata production, exchange and dissemination;
§ European statistics might not get documented appropriately by the ESS reference metadata needed;
§ The technical interoperability for further integration of metadata systems in Eurostat and Member States must be tested in practice.
7. Next steps
If the version 2.0 of the ESS reference metadata standards is approved by the DIME/ITDG, the standards approval procedure can be launched. The standards approval procedure needs to be defined (i.e. the procedures used so far or, if available, the standardisation procedure set up by the ESSnet). During the second half of 2015, the European and national metadata systems should be modified in order to accommodate the version 2.0 of the standards. The reporting according to the new versions of the standards is scheduled from January 2016.
Depending on the target date for a full standardisation of the ESS reference metadata a detailed timetable for the implementation of the standards for a full coverage of the statistical processes related to European statistics will be drawn up accordingly.
The changeover to the version 2.0 of the ESS reference metadata standards and the implementation of the standards for a full coverage of the statistical processes should be presented to the ESSC.
The Task Force on metadata standards implementation has also identified the need for training, not only on the tools for reporting on metadata and quality, but also how to report good metadata on quality (i.e. content) and such a course should be set-up already in 2015.
Eurostat will continue to work in 2015 with a few countries for improving the interoperability of the national and European metadata systems.
25
Annex 1
The revised version of ESMS (ESMS 2.0)Concept Name / Concept Name / Concept Name
1 / Contact / 7 / Confidentiality / 15 / Accessibility and clarity / The concepts “Dissemination format” and “Accessibility of documentation” are merged. The concept is renamed by "Accessibility and clarity" and reordered (former item 10 and item 11).