4.1 Metadata system(s)
The subject of this case study is IMD SDMS. This system provides the complete cycle of statistical data production processes for business statistics surveys.
The creation of the Core Meta data base module - Scheme 3, with fundamental models of structure of Micro data/Macro data was the primary task of the IMD SDMS development. Meta data base is linked at database structure model level with Micro data base and Macro data base see figure 2. Correctly and carefully planned databases structure model design is the basis for successful further system development and implementation.
All survey values from questionnaires are stored in Micro data base and each value has relation to cell (from Meta data base), which describes value meaning. Also each value in Micro data base has additional information about respondent, which gives current value and time period. The same situation is in Macro data base, where aggregated values are stored. Each aggregated value has reference to cell (from Meta data base), reference to each value aggregation conditions (from Meta data base) and correspondent time period.
Meta data base module contains following main applications:
• Description of statistical questionnaire;
• Description of questionnaire version;
• Description of indicators and attributes of statistical questionnaire;
• Description of content of statistical questionnaire chapters;
• Maintenance of validation rules of statistical questionnaire;
• Description of derived variables;
• Description of aggregation conditions of statistical questionnaire;
• Description of output tables conditions of statistical questionnaire;
• Grouping of classifications records;
• Common Meta data base data browsing;
• Applications of data Import/Export/Impute;
• Maintenance of Data electronic archiving system;
Description of electronic questionnaire for electronic data collection through
CSB WEB page. This is not a application, but the special medium of preparing
WEB survey using predefined before metadata variable through MS Word
and special tools provided by IMD SDMS;
Specially trained personnel (4 persons) of the Statistical Meta Information Maintenance Section under IT Department operate the Meta data base module. They have rights to perform Meta data entry, updating, changing and are responsible for accurateness of Meta data. It is very important, that Meta data entered into Meta data base are carefully checked and corrected, because these Meta data are used for automatic generation of data entry applications, validation, aggregation, reports preparation procedures as well as during data conversion for OLAP and PC-AXIS needs.
CSB of Latvia is using Integrated Metadata Driven Statistical Data Management System, which covers a part of GSBPM, and is intended as a system for processing of business statistics. A system with similar principles is being developed for social statistics now at the organization.
4.2 Costs and Benefits
System was launched in production in August 2002, and held 25 different surveys metadata descriptions at starting point.
Successful implementation formed basis for the CSB regional restructuring, which has been implemented within the period of two years from 2003 to 2004. Five Data Collection and processing centres replaced previously existing 26 Statistical Regional offices an city Riga office thus taking on responsibility for overall data collection and editing and decreasing amount of necessary statisticians working with data collection and editing from 180 to 115.
4.3 Implementation strategy
The project was implemented with a step-wise approach. From 1997 - 1999 CSB experts in cooperation with experts contracted from PricewaterhouseCoopers were prepared General Technical Requirements for the project "Modernisation of CSB - Data Management System". Technical Specification embodied key technical and functional requirements for the new system where statistical Meta data should be used as the key element in statistical data processing. A lot of additional requirements appeared within the process of development
The main business and information technology (IT) improvement objectives that the CSB intends to achieve as the result of project have been identified and are further described.
Using modern IT solutions:
• Increase efficiency of the main process at CSB, production of statistical information;
• Increase the quality of the statistical information produced;
• Improve processes of statistical data analysis;
• Modernise and increase the quality of data dissemination;
• Avoid hard code programming via standardisation of procedures and use of Meta data within the statistical data processing