UNITED NATIONS

STATISTICAL COMMISSION and ECONOMIC COMMISSION FOR EUROPE
CONFERENCE OF EUROPEAN STATISTICIANS
Forty-third plenary session
(Geneva, 12-15 June 1995)
CES/1995/R.7
14 December 1994

NOVEMBER 1994 WORK SESSION ON STATISTICAL METADATA

Note prepared by the secretariat

1. A work session on statistical metadatadata was held in Geneva, from 22-25 November 1994. It was attended by participants from Canada, Croatia, Czech Republic, Denmark, Finland, Germany, Greece, Hungary, Italy, Norway, Poland, Portugal, Romania, Slovenia, Slovakia, Sweden, The former Yugoslav Republic of Macedonia, United Kingdom and the United States. The European Union was represented by Eurostat. Representatives of the United Nations Statistical Division (UNSTAT) and the Organization for Economic Cooperation and Development (OECD) were also present. A representative of World Systems (Europe) Ltd. attended as an observer.

2. The provisional agenda was adopted.

3. Mr. Bo Sundgren (Sweden) was elected Chair, and Mr. Dusan Prazenka (Slovakia) was elected Vice-Chair.

4. The following substantive topics were discussed at the meeting:

Recommendations for future work are given below but other conclusions which the participants reached at the meeting on each of the above topics are reproduced (in English only) in the Annex of this note.

5. The topics were discussed on the basis of papers and/or demonstrations prepared by Canada (3), Czech Republic, Finland, Germany, Poland, Poland in cooperation with Czech Republic, Portugal, Romania, Slovakia (2), Slovenia, Sweden, United States (2), Eurostat (2), OECD, UNSTAT and World Systems (Europe) Ltd. (3).

6. The participants recommended that future work be undertaken by the Conference in the field of statistical metadata. To assist the participants in formulating their recommendations, the secretariat informed them of the goals that the Conference of European Statisticians had identified for work in this field, and distributed the portion of the integrated presentation of international statistical work programmes in the ECE region relating to the project on statistical metadata.

7. The participants recommended that a further work session be convened in 1996/97. Germany was asked by the participants to investigate whether they would be able to host the next meeting. It was agreed that contributions to the next Work Session should be submitted to the secretariat by the end of 1995. They recommended, therefore, that the following text be included in the programme of work of the Conference of European Statisticians:

2.3 Dissemination and interchange of statistical information

Activities of ECE

ANNEX

Other conclusions reached at the work session on the substantive items of the agenda

a) REPORTS ON PROGRESS AND EXPERIENCES IN THE DEVELOPMENT OF STATISTICAL METATINFORMATION SYSTEMS

Documentation: Working Papers by Canada, Finland, Portugal, United States and OECD.

b) PROBLEMS OF TRANSITION OF STATISTICAL METAINFORMATION SYSTEMS DUE TO METHODOLOGICAL, TECHNOLOGICAL OR POLITICAL REORGANIZATION

Documentation: Working Papers by Czech Republic, Romania, Slovakia and Slovenia.

1. The Work Session was informed about progress in the development and implementation of METIS in NSOs. There was general agreement that to consider this issue separately for developed countries and countries undergoing transition is not relevant any longer since all substantial problems in this area are more or less similar in both groups of countries. The work session noted that the further introduction of decentralized client/servers architecture and consequent needs to integrate management of statistical data forced statistical offices more than ever before to develop and quickly implement cost-effective and efficient SMSs. It noted the growing practical experience of countries in this field.

2. The growing role of marketing of statistical data in many statistical agencies and the important function of metadata in this process was reported by many participants. It was recommended to consider in future the development and implementation of those metadata holdings that deal especially with the dissemination of statistical information. In this connection, the necessity of a user-friendly access was mentioned.

3. The Work Session was informed about the metainformation project launched by OECD for its short-term economic indicators. The aim of this project is to unify the description of indicators collected from NSOs from both OECD members and other countries. Bearing in mind the importance of international economic analysis conducted by OECD and its possible influence on countries supplying data in the field of metainformation, the Work Session encouraged OECD to accelerate the development and implementation of this project.

4. The discussion showed different experiences and views about the possible use of administrative files in the design and implementation of SMSs. Some participants mentioned rather negative experiences caused, for example, by low quality of outside data for statistical purposes, missing metadata and the necessity of proper training of outside suppliers. Other countries reported rather enthusiastically, stressing cost savings, reduction of responsibility of statistical agencies for such collection, and relatively reliable response rate.

c) REPORT ON METADATA PROJECTS IN EUROSTAT

Documentation: Working Papers by Eurostat (2).

5. The Work Session was informed on the plans and major issues developing metadata in EUROSTAT. The metadata concept should be mainly macro-level oriented. The Work Session noted that EUROSTAT's effort faces some challenges which could make this task somewhat complicated, for example: the flow of standardized metadata from national statistical agencies to Eurostat, harmonization of the use of coding, concepts and classifications, the integration of metadata with nomenclature systems and multi-lingual thesauri.

6. It was also pointed out that the project DSIS (Distributed Statistical Information Services) which is organized as a framework for co-ordinating development of European Statistical Information Services also includes the development of an agreed stock of metadata as part of the common reference environment.

7. The Work Session was also informed about the work on the newly established research programme DOSIS (Development of Statistical Information Systems). The programme of work will be based on various stages of the statistical production process, that is, data collection, data processing, analysis and modelling. There will be no special section devoted to statistical metadata; this issue will form an important basis within several possible research topics.

d) GUIDELINES FOR THE STATISTICAL METAINFORMATION SYSTEMS MODELLING AND DEVELOPMENT

Documentation: Working Paper by Sweden.

8. The Work Session discussed the methodological document on the design and implementation of SMSs. Participants agreed to consider it as a final version and recommended that it be reproduced and distributed in 1995 as a guide produced under the auspices of the Conference of European Statisticians. The author was requested to deliver the final document in hard copy and on diskette to the ECE secretariat by the end of February 1995. The title of the guide shall read: "Guidelines on the Design and Implementation of Statistical Metainformation Systems".

e) GENERIC MODEL FOR STANDARDIZATION OF STATISTICAL INDICATORS

Documentation: Common Working Paper by Czech Republic and Poland.

9. The Work Session considered the final version of the generic model for the description of statistical indicators. The new title: "Standardization of Statistical Indicators and Metadata" for this document was adopted. The discussion showed that this model can be used for the description at the micro-level as well as at the macro-level of statistical indicators. The Work Session recommended that it be reproduced and distributed in 1995 as a guide produced under the auspices of the Conference of European Statisticians.

10. Several participants requested an explanation of the relationship between the presented model and the GESMES standard message developed in the framework of UN/EDIFACT. It was clarified that the role of GESMES is to give a standard format for the exchange of macro-data in the form of multidimensional statistical tables and time series. The scope of the presented generic model for the description of statistical indicators is wider. It was agreed that this material can be considered as self-sustainable.

f) COMMON TERMINOLOGY FOR STATISTICAL METAINFORMATION SYSTEMS

Documentation: Working Paper by Slovakia.

11. An updated version of the Common Terminology for Statistical Metainformation Systems was presented to the Work Session. In addition to the 1993 version, the document was newly formatted, the definitions and cross-references were revised and the glossary was completed with new terms.

12. The Work Session stressed the importance of this material for further work in SMSs. To accelerate its finalization, a special expert group was created with the goal to review and finally edit all terms. The group will work on a correspondence basis and will finalize its work by the end of 1995. Representatives of Canada, Czech Republic, Poland, Slovakia, United States, UNSTAT and World Systems (Europe) Ltd. volunteered to take part in this work.

13. Furthermore, it was suggested to study possibilities to disseminate the glossary electronically. In this connection, the form of hypertext was mentioned.

g) INVENTORY OF STANDARDS APPLICABLE IN STATISTICS

Documentation: Working Paper by Canada.

14. The Work Session considered possibilities of the further development of the Inventory of International Standards for the Representation of Statistical Data and Metadata (issue published in 1993). The representative of Canada proposed to incorporate in the Inventory the ISO Information Resource Dictionary System (ISO/IRDS) family of standards. The Work Session requested the ECE Secretariat to update the Inventory in this respect. In addition, the Canadian representative explained the role of ISO/IRDS framework and its possible influence on the organization of information holdings.

h) DESIGN OF METADATA AND INFORMATION HOLDINGS

Documentation: Working Papers by Canada and Poland.

15. The presentations showed that particularly the following types of metainformation holdings can be identified: catalogues, dictionaries, directories, thesaurus and registers. The implementation of these holdings differs depending on architecture of SIS, software tools used, etc..

16. The discussion revealed that the growing number of requests to integrate classifications and nomenclatures calls for complicated solutions through correspondence tables. Another frequently mentioned problem was the need for multi-lingual interpretations.

17. It was pointed out that the variety in the organization of statistical databases is accompanied by a certain diversity in the way in which metadata holdings are linked to databases. Since statistical data in client-server architecture are often stored in distributed databases, the metadatabase should create an integrated platform for the orientation of users.

18. A software tool developed for the dissemination of the Inventory of Computerized Databases and United Nations Statistical Yearbook on CD-ROM was presented to the Work Session. Participants noted with interest this possibility of dissemination of statistical data and metadata, and questions such as standardization and unification of classifications and nomenclatures, validity and quality of data in relation to metadata and interpretation of footnotes were discussed in connection with the further development of these electronic products.

i) MANAGEMENT ASPECTS OF STATISTICAL METAINFORMATION SYSTEMS

Documentation: Working Papers by United States, Germany and World Systems (Europe) Ltd.

19. The presentations showed diverse organization of maintenance and consequent management of metadata in distributed database environments. While in some countries the maintenance of metadata holdings is centrally managed, in others responsibility is given to administrators of decentralized databases.

20. There was common agreement that a well-organized and efficiently standardized SMS should be employed as a tool to reduce the complexity and diversity of the whole statistical information system. However, many participants mentioned that proper motivation and recognition of the work done by staff in statistical agencies for the implementation of such tools is often a challenging task.

21. There was also general agreement that, when designing such systems, the reality of statistical production should be taken into consideration. Some participants mentioned that, although standardization is needed in certain stages, the alternatives should be available to efficiently satisfy a variety of user demands. EMMA, the software developed by World Systems (Europe) Ltd., was presented as one of the possible supports for the metatinformation management architecture.

22. It was also mentioned that the necessity of the multi-professional approach when designing and implementing SMSs may introduce some new managerial aspects in statistical agencies.

j) EVALUATION CRITERIA FOR STATISTICAL METAINFORMATION SYSTEMS

Documentation: Working Paper by World Systems (Europe) Ltd.

23. Based on the proposal presented by World Systems (Europe) Ltd., the following possible phases of the evaluation process were discussed: determination of requirements, establishment of criteria, development of the cost/benefit model, performance of the evaluation process and examination of results. Some participants mentioned that, when evaluating SMSs, the efficiency and design of implementation of both software tools and metainformation holdings should be considered.

24. A common opinion was expressed that there is a lack of methodological materials for the evaluation of SMSs, and therefore the further development of such materials is highly desirable. There was general agreement that evaluation criteria should be applied already in the process of design of SMSs and not only in the process of their implementation. It was also mentioned that the development of generally valid evaluation criteria would be scarcely possible.

k) FORMAL METHODS FOR THE DESCRIPTION OF METADATA

Documentation: Working Paper by World Systems (Europe) Ltd.

25. The Work Session was informed about the progress in the implementation of World Systems' projects on the Statistical Metainformation Language Environment (SMILE) and the Expert Interface to Statistical Information Systems (EISI). These projects were specified using formal languages -- Vienna Development Methodology and Z Specification Language -- in their development of software tools for SMSs. The functional model is based on methods described in the "Standardization of Statistical Indicators and Metadata".

l) OTHER SUBJECTS

26. The Work Session asked the representative of Canada to make the work of the project on statistical metadata known to the ISO/IEC JTC1/SC21 WG3 Rapporteur Group on the Information Resource Dictionary Systems (ISO/IRDS) through his role as a member of the Canadian Advisory Committee to ISO/IEC JTC1/SC21 WG3.

ANAGEMENT ASPECTS OF STATISTICAL METAINFORMATION SYSTEMS

DocOwO,28{ ]*s9_^FF3~ k  * 0 v }  |  F . z AHM%sbbnt_Dgma7=@ ]?1n(VY3;B;A/$xi X )!}!!"m"m""""I###5$$$)%<%r%x%%%%H&&&-'-''''(c((( ))b))*U****M+d+j+++++<,,,4-----:.@....*/u//0`0`0001Y111G2222 3&3U3[33394444475556g6m66 77)7g7m7777H888<9999::V:\::::0;~;;;<j<<<<2==== >^>d>>?U???D@@@@@@KA[AaAABVBVBBBB