The MetaNet network of excellence sought to harmonise and synthesise the numerous statistical metadata developments taking place by providing a much-needed focal point, bringing together theoretical and practical expertise spanning the entire spectrum of statistics.
- to develop proposals for standards in the methodology used for describing statistical metadata and statistical information systems
- to develop proposals for recommendations on the metadata objects in a common conceptual model of statistical metadata
- to disseminate these proposed standards to the relevant user communities and standards bodies
- to interact with relevant FP5 projects on the development and agreement of these proposals, and to advise on methods of achieving coherence of approach in the field of metadata for statistical information systems
- to integrate the different views of metadata into one model and bring together these different perspectives
The decision to form MetaNet was based on the belief that interest in statistical metadata was wide-ranging and that the time was right for interaction between all circles of interest. This belief was substantiated, with members and associates introducing important insights from 2 international organisations plus 23 national organisations, spanning 14 countries. The network thus served as a microcosm of metadata specialists who, despite working in different contexts, with different kinds of data and having different goals, were nevertheless eager to understand each other's point of view and work towards common goals. The integration of such experts under the umbrella of MetaNet in itself can be considered an important 'value added' result of the network.
The method of consolidation ensured that the range of contributions considered by the network, whilst not exhaustive, reflected the diverse perspectives of everyone involved. The conceptual framework developed supported the understanding of the commonalities and differences of data/metadata models from a statistical point of view. This constituted a significant step forward in orienting the different aspects of metadata to each other and represents a potential future standard. An important finding of MetaNet was that given the heterogeneity of statistical metadata, the adoption of different specific metadata models for particular functions and domains of statistics is something of a necessity. However, all such models now enjoy a common conceptual platform.
The project considered how best to develop and implement statistical metadata infrastructures supporting the production and usage of statistics. The value of the results of this particular undertaking has been widely affirmed by the statistical metadata community in Europe and beyond, significantly those players concerned with the day-to-day practicalities of implementation. This work therefore represents a notable process innovation.
The survey conducted importantly confronted some of the main findings of the other work groups with a broader user group and obtained some very interesting results, which should inform the future research agenda. Significantly, the finding that there was no clear consensus on a definition of statistical metadata or a common understanding of the borderlines suggests that perhaps the concept is becoming too broad and it might be more useful to focus on specific functions. Moreover, the gap between some specific metadata models and tools in development and the integrated and practical solutions asked for by subject matter specialists clearly points to the need for a stronger involvement of the latter group in the future development of metadata solutions.
Publications arising included: an inventory of current metadata methodology and tools; a report on metadata concepts and how these can be used in practice; a reference book for metadata standards and methodology; and a training information manual for the adoption of metadata standards and systems.
MetaNet enabled the UK Data Archive to engage in discussion and exchange knowledge and information with a range of data and metadata producers both inside and outside academia. Particularly, we shared our expertise in implementing the DDI within Nesstar.