Abstract: Statistical meta-knowledge and -data requirements are described in this paper and the desired content of metadata in statistical systems and their components of statistical metadata systems are outlined. Finally, the needs for further efforts in developing components missing and for consolidating research results are emphasized.
Abstract: This paper outlines some of the requirements for statistical metainformation systems from the viewpoint of potential users. Users can be staff in statistical offices who provide statistical information to the public or external clients. Keeping user requirements in focus will increase the chance to invest development resources where they bring the greatest payoff. The nature and variety of such requirements are illustrated using examples from Statistics Canada.
Abstract: A conceptual framework is presented, covering pragmatical, semantical, and syntactical aspects of statistical metainformation systems. Examples from on-going projects in some statistical offices are used as illustrations.
Abstract: Data consist of more than merely the numbers representing the magnitudes of attributes of the objects being investigated. The other component of data is the meaning of the numbers – the context in which they arose – the “metadata”. A distinction is made between context-specific metadata and context-free metadata. The latter is illustrated in depth with the example of measurement scale. Brief discussions are given of general issues relating to metadata research.
Abstract: The concern of this paper is with the much neglected area of footnotes to statistical tables. The role of footnotes in a statistical metainformation system is outlined with particular emphasis being given to user (as opposed to producer) needs. Footnotes are placed within the user's information context and their role and form analysed. The analysis ia grounded in a study of the actual usage of footnotes in a major statistical publication and a functional taxonomy derived. The state of footnotes is argued to be in serious disarray. In spite of their obvious importance the subject is bereft of study. The…responsibilities of Statistical Offices vis-à-vis their use is argued to require explicit guidelines as to good practice. The increasing availability of series in electronic form without conventional footnotes to warn of comparability problems may be considered irresponsible.
Abstract: This paper describes the “Modelling Metadata” project, which is one of the projects in EUROSTAT's DOSES (Development of Statistical Expert Systems) program. The objective of the project is to develop a general methodology for the study of metadata. The ultimate aim of our research into metadata is to establish a framework for a computerized system that will support metadata-based statistical data processing. We describe the methodology used to design the system, and review the characteristics of an implemented prototype. A central characteristic of the methodology – which is reflected in the name chosen for the project – is the…formalization of metadata by means of conceptual models.
Abstract: Focusing on metadata referring to the meaning of statistical data in the field of Official Statistics, a data model is proposed supporting the semantic integration of different data sources. To this end, the different empirical languages of data sources are re-expressed in terms of a common theoretical concept language. Moreover, based on this theoretical data language a table specification language for describing the structure of target tables is presented in an approach to obviate the need of stating procedural derivations explicitly. Since statistical tables are considered as the elementary data structure of the proposed approach, the corresponding computation model for…the process of table generation is outlined, too. Finally, the relation of this work to other metadata research, in particular to the DOSES project “Modelling Metadata”, is discussed, highlighting perhaps some of the most important yet still untackled problems of metadata research.
Abstract: A present project at Statistics Sweden to implement a statistical metainformation system is described. The system is to support documentation work for statistical production systems and is based on the SCBDOK structure as specified by profs Rosén and Sundgren at Statistics Sweden. Emphasis is on the reuse of these documentations to create metainformation systems with other purposes.