Case study: Statistical Office of the Republic of Slovenia

 1.1.1. Corporate strategy and guidelines for the preparation and dissemination of metadata across the organisation

A centralised, corporate metadata repository is implemented at SORS. It includes metadata about surveys, publications, statistical terminology, classifications and nomenclatures, advance release calendar. For each statistical survey methodological explanations have been developed and are available on SORS's website (in English as well).
This year the project to extend the metadata repository has been developed. The repository will include questionnaires, variables and connections to databases. The next step will be to connect those two parts with the repository for reference metadata, which will have to be built as soon as possible.
In the centralised metadata repository metadata will be reusable among different surveys and among different statistical domains. They will be disseminated through our website.
Due to the Memorandum of Understanding that was signed by different international organisations, SORS decided to consider SDMX standards as the basis for dissemination of metadata to international organisations. At the time being SORS is involved in the project SODI  and is testing the "push" method for sending data and metadata to Eurostat.
Implementation of SDMX standards is a technical and a content problem. To be as systematic and effective as possible, a small group of four people has been established at SORS. The main goals of the group are: 1. to monitor the development of the SDMX, 2. to collaborate in different Eurostat activities in the field of development and implementation of the SDMX, 3. to prepare suggestions on how to take steps to implement the SDMX, 4. to acquaint the wider interested public with the SDMX. For SORS's staff metadata and statistical glossaries are available on the intranet. The metadata Common Vocabulary (MCV) is also in the preoperational phase. Because it is a part of SDMX standards, SORS puts great attention to the usage of common methodology for metadata in all statistical domains. Here is to be mentioned the internal SORS's Terminological Group, which is: 1. authorized to give the opinion about initiatives of new statistical terms within the scope of statistical terminology, 2. responsible for the regulation of the terminological database and its maintenance.

1.1.2. Existence of corporate database / glossaries to promote the use of standard concepts and definitions

It is the usual practice that statistical information is accompanied by standardised metadata, available on the website for all users. There is a responsible officer in SORS who checks that metadata are available, and monitors metadata documentation.
Methodologies are available on the website and changes are communicated to users and discussed in the statistical advisory committees which include also users of statistical information. SORS has published a number of comprehensive quality reports, consistent with the ESS quality criteria, and plans to add more every year.

1.1.3. Processes to ensure that sampling and non-sampling errors are measured, systematically documented and information made available to users for all key statistical outputs

The process to ensure that sample and non-sample errors are measured is very well used within those statistical surveys which in the recent years have gone through major technical revisions. For these surveys standard quality indicators, among which are estimations for sample and non-sample errors, are calculated automatically within the statistical process. Indicators are available in Standard Quality Reports published on SORS's website (from September 2008 on available also in English).

1.1.4. Metadata are documented according to standardised metadata systems

Detailed metadata in English can be accessed from the SORS website for over 160 subjects which may also be grouped by the user into 30 fields of statistics. Although the metadata items vary by subject detailed metadata are generally available for:

-         purpose of the survey; 
-         observation unit; 
-         period of observation; 
-         sources, coverage and method of data collection; 
-         definitions and explanations; 
-         where data are published; 
-         date when metadata were last updated; 
-         name of person who prepared the metadata. 

A Guide through Statistics glossary on the SORS website provides key definitions from statistical surveys and provides an overview of statistical standards used by individual fields of statistics. The glossary does not provide the source for each definition.

1.1.5. Recommendations / guidelines for the treatment of time series breaks

SORS and the authorized producers in most cases report that they clearly notice the breaks in time series caused by different reasons.
Although national statistics doesn't have special written procedures on the treatment of time series breaks, SORS through the Editorial Board centrally coordinates the publication of the time series breaks in a standardized way. 

1.1.6. Recommendations / guidelines for the presentation of time series - including seasonally adjusted data

Since 2005, internal recommendations for the presentation of indices (WDA, SA,...) have been used at publications and SI-STAT database in the field of short term statistics. These guidelines will be upgraded and the use for the publications in 2010 will be obligatory for all time series. Besides, special guidelines were developed for the publication of time series in the case of Euro changeover.

1.1.7. Within the MEDIUM-term programme of statistical surveys: 2008-2012 we read:

Current Medium-term Programme of Statistical Surveys 2008-2012 is the third such document in the history of Slovene national statistics. Knowledge and experience obtained in the past period were the basis for preparing the objectives and activities that are part of the Medium-term Programme of Statistical Surveys 2008-2012. Some of the tasks that link both medium-term periods (the other was (2003-2007) are: 1. active presidency of the EU Council Working Group on Statistics during Slovenia's presidency of the EU, 2. Slovenia's preparations for membership of the OECD, 3. preparations for the Census of Population and Housing, which will be conducted for the first time fully on the basis of administrative sources and 4. development projects and temporal and financial analysis of the burden of reporting units and producers of national statistics, which takes place within the Commission and will contribute to further reduction of the reporting burden and costs while achieving the same or even better quality of statistical outputs and services.
A new feature of the medium-term programme is nine strategies that are the framework for quality improvement of already known data requests and for creating the conditions for recognising new contents that will be needed in the future. Strategic activities will be planned for two years in action plans, while their implementation will be monitored with annual reports.
The nine strategies are: 
-         Strategy for the further development of national statistics in Slovenia; 
-         Strategy in the field of further reduction of administrative burdens;
-         Strategy of quality;
-         Strategy of data protection;
-         Strategy of cooperation with reporting units and reduction of burdens;
-         Strategy of dissemination and communication with users;
-         Strategy of information technology update;
-         Strategy of human resources management;
-         Strategy of financial resources management,

Two of the most important main objectives of the medium-term programme of national statistics for the forthcoming five-year period are to provide quality and user-friendly national statistics, and to improve the flexibility of the response of national statistics to the needs and requests of users by using new sources of data. At the same time all institutions in the Slovene national statistics system set themselves a goal to strengthen the trust in national statistics, not only on the part of users but to the same extent also on the part of statistical data providers - including the efforts to reduce their burden when reporting data to producers of national statistics. In the next five years data collection will be rationalized with even more intensive use of administrative sources and new innovative approaches to achieving greater efficiency will be designed.

1.2.1. Slovenia has a centralised, functionally de-concentrated institutional statistical system

SORS as the main executive and coordinating authority of the national statistics with assistance of authorized producers of national statistics provides relevant statistical information to national and international society and officially delivers data to Eurostat. Moreover, SORS provides to the public administration bodies and organizations, the economy and general public data that are of good quality, timely and comparable in time, space and internationally, on the status and trends in the economic, demographic and social fields, as well as in the fields of environment and natural resources. At the same time aims at providing data for dissemination at minimal costs. By respecting statistical confidentiality, SORS builds users and respondents' trust and its professional independence.
A register-oriented system: Slovenia has a relatively developed infrastructure of public evidences and administrative registers. The development of administrative registers and their interoperability is a strategic orientation of the country. Multi-use of registers and their outstanding role in the implementation of e-services are of key importance. One of the most important users of administrative sources of data is the national statistical system, which can be considered as a register-oriented system. Article 28 of the NSA states that administrative and other national authorities, public servants and holders of public offices, as well as authorized producers which collect statistics in their own fields of work and who are not included in the programme of statistical surveys shall define the methodological bases of their own statistical surveys after having obtained prior opinion of the Office.

Methodologies are available on the website and changes are communicated to users and discussed in the statistical advisory committees which include also users of statistical information.
Metadata (Methodological Explanations): It is the usual practice that statistical information is accompanied by standardized metadata, available on the website for all users. There is a responsible officer in SORS who checks that metadata are available, and monitors metadata documentation. Detailed metadata in English can be accessed from the SORS website for over 160 subjects who may also be grouped by the user into 30 fields of statistics. Although the metadata items vary by subject detailed metadata are generally available for:
-         purpose of the survey; 
-         observation unit; 
-         period of observation; 
-         sources, coverage and method of data collection; 
-         definitions and explanations; 
-         where data are published; 
-         date when metadata were last updated; 
-         name of person who prepared the metadata. 

A Guide through Statistics glossary on the SORS website provides key definitions from statistical surveys and provides an overview of statistical standards used by individual fields of statistics*. The glossary does not provide the source for each definition.

SORS' release calendar is available from the home page of the website from an interactive facility. The site enables users to retrieve release information by field of statistics, subject and series for a time span specified by the user. The facility also allows the user to view all publications released on the current day, week or month. Finally, users may select fields of statistics for which they wish to receive notifications about new releases. The calendar is regularly reviewed by senior management. To cater for releases well into the future SORS maintains a "release not later than" category - releases may be earlier than this (though in practice are firmed up nearer the time) but never later. Any divergence from the dissemination time schedule is publicised in advance, explained and a new release date set. The problem of divergence is well arranged by SORS's internal regulation which gives detailed explanations about the planning procedure for the release calendar.  The regulation defines that in case of unexpected events, for which statistical data can not be provided on time (not even estimation!), although publication was planned and its release date confirmed, SORS on its website on the scheduled day publishes special information and notifies users with short explanation by e-mail as well. This information is removed from the website, after confirmation of the next release date. However, these events are extremely rare (once or twice per year). 

1.2.2. Priorities of the National Statistics in 2009  
 

1.2.2.1. Total quality management at SORS

In 2009 the SORS priorities will be set considerably wider, and will cover: 
-         activities to promote the cooperation of enterprises in data reporting; 
-         concern for the consistent monitoring and analysis of response burden; 
-         implementation of the process of standardising questionnaires in statistical surveys; 
-         preparation of documentation on the consistency and coherence of statistical data from various surveys; 
-         preparation of internal rules and procedures on revisions of published statistical data; 
-         establishing internal rules and procedures for keeping documentation for the time series breaks.

In line with the adopted action plan for implementing the quality strategy, great attention will be focused on the preparation of internal methodological manuals (i.e. textbooks, handbooks, presentations, etc.) for individual parts of the statistical process. At the same time standard and annual quality reports will be prepared.By providing the necessary technical and expert support for establishing statistical methodology, SORS attempts to increase interest in implementing the quality standards of the European Statistical System among authorised producers of national statistics. As part of the assessment of the compliance with the principles of the Code among authorised producers, for 2009 SORS is planning to carry out a special series of training sessions for authorised producers. In terms of content, apart from the preparation of standard and annual quality reports, the training will cover the formulation of methodological explanations and a release calendar. At the same time a short course is being planned on imputing missing values, data editing in statistical surveys, on sampling methods and data dissemination.

1.2.2.2 SORS infrastructural project

Completion of the project aimed at modernisation of statistical processing for business statistics (ISIS information system) at the end of 2008 means that SORS will face a major change in the method of work, and this will be visible in 2009. In addition to optimising the current procedures, the changes will also affect work, which will henceforth be based on metadata driven processes; individual processes will be carried out in a standardised way, which will also contribute to less time spent.
Changes to procedures affect all areas of statistical process:
from the preparation of statistical survey through to the introduction of a common base of variables, questions and questionnaires,
data collection through the introduction of e-reporting, optical reading of printed questionnaires, the implementation of communication types and a common directory of business that will be based exclusively on the statistical business register,
data processing through the inclusion of data from various sources, simplified metadata processing and so forth.
At the end of the project and its implementation in practice in 2009, SORS must carry out several small organisational and procedural changes to support the new infrastructural purposes - i.e. support for enterprises (reporting units) in implementing e-reporting, establishing a call centre and so forth. The transition of statistical surveys to the new system will be gradual for enterprises and SORS. 

1.2.2.3 Establishing a new statistical business register within the ISIS framework

In 2009 upgrading will be completed on the statistical business register, which will include the Register of Enterprise Groups. With this project SORS will succeed in responding to the majority of requirements of the EU regulation on business registers and also the requirements of internal SORS users, i.e. our employees. Indeed for business entities the register will enable the central keeping of data, which through the use of appropriate statistical methods will be conducted on the basis of a combination of administrative and statistical data.

1.2.2.4 Further development of cooperation with reporting units

While it continues its efforts to reduce the response burden, SORS is also seeking appropriate technical methods for reporting data. In addition to the aforementioned e-reporting, in 2008 we started introducing internet telephony. In addition to automatic dialling, one major task will be establishing an Information base for reporting units, accessible on the SORS website. This will contribute to better quality communication with reporting units, which is also in line with the medium-term strategy 2008-2012.

1.2.2.5 Concern for the reduction of response burden and cost effectiveness 

Monitoring and analysing the burdening of reporting units and producers of national statistics
Based on the Action plan to reduce administrative burdens and costs, which was published on 24 January 2007 by the European Commission, on the Commission's request SORS estimated the time and costs of reporting units and producers of national statistics for statistical surveys conducted on the legal basis determined for this estimate by the Commission. The estimate for 2005-2007 was carried out on the basis of the internationally adopted methodology of the Standard Cost Model. By the end of October 2008 we will also have completed measurement of the burden for all tasks of the annual statistical programme for which SORS is responsible. The findings will be an important basis for the further reduction of response burden and rationalisation of the work of SORS.

1.2.2.6 Classifications

Introducing KLASIUS into procedures for accreditation of programmes and catalogues and into data collections and statistical surveys

In 2009 SORS expects the competent ministries, institutions and other bodies, in line with the Agreement on cooperation in implementing the classification system for education and training, to implement KLASIUS in procedures of accepting the accreditation of programmes and catalogues for national vocational qualifications, and to ensure that the existing programmes will also be furnished with KLASIUS codes. This will enable SORS to set up in 2009 a new base - a statistical collection of educational activities/outcomes, which will form the basis for drawing up detailed coded KLASIUS lists, and e-browser and other aids. These activities will be followed by the second phase of introducing KLASIUS, specifically in data collections in which the primary unit is a person, education or inclusion in education as the characteristic of a person - starting with records on labour market statistics, M forms  and records or databases from the Employment Service of Slovenia.

The current metadata system was gradually developed, starting in 1997. 

4.2.1 
It started with a "Modernisation and development of the statistical information system in Slovenia", Feasibility Study on the Architecture of Information Systems and Related Equipment Issues. 
The study was carried out in the period of February - September 1997. A number of short term missions to SORS by experts from Statistics Sweden took place. Among the main conclusions of the study were: 
-         SORS has an excellent potential for developing a modern, register-based statistical system, based upon administrative sources in combination with sample surveys (which need to be designed in a more optimal way than at present). However, there are many demanding tasks to be tackled within a relatively short time period and with quite limited resources. In this situation there is an urgent need for focus and systematic planning in the development work. 
-         One way to obtain better focus in the development of the systems for statistics production in Slovenia is to specify very precise and concrete target architecture for the development, and to formulate a strategy for implementing this architecture step by step, in a systematic way. As a matter of fact, this approach has been proposed by SORS itself. Both top management of SORS and other staff members have expressed their sincere interest in establishing an "ideal" target architecture and a systematic implementation plan. 
-         The information system architecture for a statistical office should cover a number of different information systems types and their relations to each other: registers, survey processing systems (primary systems), analytical systems (secondary systems), and metainformation systems. 
-         It was further recommended, that the first step in the proposed architecture should be the building of a classification database. The prototype was presented at the board of director general as of 1 March 2000 and put in production in November 2000.

4.2.2 
Within the StatCop98 project, the component 4.1: Development of conceptual, technical and software solutions of common (infrastructure) importance had the following goals: 
-         creating the concept of a statistical data warehouse with special emphasis on common functions and metadata as well as its testing on a pilot project; 
-         specification, development and introduction of EDI tools and procedures; 
-         classification database - upgrading the existing functionality, developing software for managing the concordances; 
-         developing software for browsing classifications via internet.

Another component - 4.3: Development of databases and software solutions - aimed at an integrated process of aggregation and dissemination of data from the Census of agriculture, horticulture and viticulture 2000 (AC2000) and other agricultural statistics (AGRISTAT). Within these two components, the basic common functions in the context of statistical data warehouse were defined according to Sundgren (Sundgren 1997): "Statistical metadata are descriptive information or documentation about statistical data, i.e. microdata, macrodata, or other metadata. Statistical metadata facilitates sharing, querying, and understanding of statistical data over the lifetime of the data". 

4.2.3 
The next project, STAT 2000 - focused on dissemination procedures. At that time, we believed that the electronic dissemination procedures at SORS are not well adapted to user needs and EU requirements. Most of the data collected by SORS are to the users available disperse, i.e. not in an integrated and comprehensive form and in addition they mostly lack the metadata. All official statistics have to be available in a uniform and user-friendly way via the Internet. Providing users with ample amounts of high-quality metadata encourages them to process and analyse official statistics in their own computer environments and to give feed back to the statistical office. The interaction between production and dissemination systems has to be organised on metadata based concept, so that the production system automatically feeds the output database, used by dissemination systems, with data and metadata. 
An integral and contemporary dissemination system should be designed to cover the national and international user's needs. The data used in the compilation of SORS's dissemination system will cover all fields of statistics as well as series of regional data indicators for which the metadata system has already been initiated in the scope of the COP 98 project. 
Metadata play a major role in dissemination of statistics, including helping users to find, understand and assess statistics in the context of their specific objectives. As standard tools and approaches to creating and managing, the metadata used for dissemination can also be used for survey design and statistical production activities. 
Sharing information and metadata requires standard solutions both for the technology and the content. 
Within the STAT 2000 project it was therefore essential to analyse and establish the underlying principles for specification and modelling of the dissemination database. 

Figure 4:  Conceptual scheme of functions to be supported within the "output" process 

It was essential to study the entire data and metadata flow as far as possible all the way from the microdatabase to allow effective and efficient process, based on guidelines and rules that facilitate both: first production of statistical data and any other future use and reuse of the data and metadata. 

Figure 5: Conceptual scheme of the data and metadata flow from the microdatabase to the end user

Despite the fact that a large part of the development tasks in the area of information services for the support of statistical production should be finished by the end of 2008, the introduction of solutions into regular production is a particular challenge, and will continue at least until 2010. The introduction of new applications and tools into general use will increase SORS's need for constant improvement of the level of IT services (providing functioning, user support, infrastructure management, etc.) in order for the second main goal of the strategy to be efficient, and that internal and external users of information services are satisfied.

The following actions plans are elaborated in *Priorities of the national statistics in 2009 (Annex 6):

-         Action plan for implementing the STRATEGY FOR FURTHER DEVELOPMENT OF NATIONAL STATISTICS IN SLOVENIA in 2009; 
-         Action plan for implementing the STRATEGY FOR FURTHER REDUCTION OF ADMINISTRATIVE BURDENS in 2009; 
-         Action plan for implementing the STRATEGY OF QUALITY in 2009; 
-         Action plan for implementing the STRATEGY OF DATA PROTECTION in 2009; 
-         Action plan for implementing the STRATEGY OF COOPERATION WITH REPORTING UNITS AND REDUCTION OF BURDENS in 2009; 
-         Action plan for implementing the STRATEGY FOR DISSEMINATION AND COMMUNICATION WITH USERS in 2009; 
-         Action plan for implementing the STRATEGY OF INFORMATION TECHNOLOGY UPDATE in 2009; 
-         Action plan for implementing the STRATEGY OF HUMAN RESOURCE MANAGEMENT in 2009; 
-         Action plan for implementing the STRATEGY OF FINANCIAL RESOURCES MANAGEMENT in 2009.

SORS is a user-oriented organisation which is organised with an emphasis on processes and the quality and protection of data provided by reporting units. Information technology (IT) must follow the mission, vision and values of SORS, and provide for the efficient, professional and economic execution of SORS's mission and tasks.

Key goals for the realisation of the tasks set forth are:
1. well computerised and efficient statistical process, supported by general user-friendly information solutions,
2. efficient and satisfied internal and external users of information services,
3. competent IT experts,
4. electronic storage and archiving.

Computerised and efficient statistical process
This will be achieved by developing and introducing general information to support the statistical process, which enables more transparent and efficient statistical processes and long-term savings in the maintenance of these solutions.
Such an approach brings new challenges from the aspect of development and introduction. Planning, development, testing and the introduction of general information systems require the use of better defined and adequately supported methodology in terms of organisation. The IT unit will start introducing established standards and development methodologies which cover the entire life-cycle of the application software (CMMI). The next important activity is improving organisation within the IT by introducing process organisation of development activities, which will enable IT experts to specialize in individual expert areas and be independent from the statistical content in their work. This will lead to a more transparent process in the development of information systems, a higher concentration of IT knowledge, and easier and more efficient allocation of human resources.
 

Consolidation and standardisation of software
SORS has consolidated its development environments in recent years and will standardise them considering individual statistical processes. Besides a long-term focusing on relational technology of Oracle databases (and in certain areas of Microsoft SQL Server) SORS has chosen NET as an application development environment, SAS as the environment for processing and analyses, and tools developed in statistical offices (Blaise, PC-Axis, Argus, etc.) for specific statistical areas (collection and arrangement of data, dissemination, protection, etc.). The use of commercial and statistical tools wherever possible and the development of our own application solutions only where they are necessary for the smooth functioning of the statistical process remains a constant guideline. The main challenge will therefore be to integrate individual technological solutions into an integrated system, which will be achieved with the improvement of interoperability between individual systems and the introduction of the principles of service oriented architecture.
A very important activity will be finalising the transfer of processing and data into a LAN server environment, with which SORS will conclude the use of its IBM central computer, and will also transfer data and processing performed on working stations. Due to the increasing importance of XML technology in the area of statistical data and metadata, SORS will start developing and introducing SDMX technological solutions.

The first step in this direction is the introduction of ITIL guidelines, which is a long-term process of constant improvement, which will enable SORS employees and external users to work smoothly and efficiently. In terms of organisation, this will bring the establishment of a modern service centre and adequate process organisation in the area of IT services.

Shared servers and dedicated ISIS servers

Figure 6: ISIS system architecture

1.
Participation in pilot projects enables less experienced employees to gain the experience necessary for independent work. Due to the similarity of the statistical process, it is very important for the IT personnel in SORS to gain experience from other statistical offices, which is ensured through participation in expert conferences and bilateral cooperation with foreign offices.

2.
Metadata usage in production tools has shown many opportunities but also a lot of new challenges.

3.
In 2009 within total quality management priorities will cover:
- activities to promote the cooperation of enterprises in data reporting;
- concern for the consistent monitoring and analysis of response burden;
- implementation of the process of standardising questionnaires in statistical surveys;
- preparation of documentation on the consistency and coherence of statistical data from various surveys;
- preparation of internal rules and procedures on revisions of published statistical data;
- establishing internal rules and procedures for keeping documentation for the time series breaks.

4.
Great attention will be focused on the preparation of internal methodological manuals (i.e. textbooks, handbooks, presentations, etc.) for individual parts of the statistical process.

5.
By providing the necessary technical and expert support for establishing statistical methodology, SORS attempts to increase interest in implementing the quality standards of the European Statistical System among authorised producers of national statistics. This will cover also the formulation of methodological explanations and a release calendar. At the same time a short course is being planned on imputing missing values, data editing in statistical surveys, on sampling methods and data dissemination.

6.
SORS has been building a new metadata driven system during the last few years. This information system is designed in a highly flexible and scalable way, based on modern technological approach (SOA), to assure easy adaptation to future changes, reusability of developed services and further development of the SOA end electronic business at SORS. Relevant actors (users) were engaged in testing in early phases of project to know the new system and to be able to start use it after the end of the project in the most efficient way.

6.1 Descriptions of the problems

6.1.1. Complexity of the project / relatively short time for implementation
The new project has been very complex and ambitious. Information system covers metadata management system, entire process of flexible and metadata driven statistical surveying, sophisticated electronic reporting system, Statistical business register and special respondent's management system, seasoned with heterogeneous technological platform and all kinds of integration demands with specific statistic tools.

6.1.2. User specifications
Consultant did prepare technical documentation during analysis and design phases which has described the future information system; however, the technical documentation in most cases was not transparent enough for users to understand the designed system in details before it was built. The main problem was »translation« from »statistical language« to »IT language« and it caused many misunderstandings.

6.1.3 Organizational adaptation of SORS
Information system like the newly developed, inevitable results in fundamental adaptation of organizations. It should enable SORS to turn paper-centric statistical surveying into paperless statistical surveying - it is hard to imagine a bigger change.
The organisation of  data process has been under process of adaptation to implement the metadata driven system successfully in its processes. Implementation of electronic reporting alone demands entirely different "service and customer-centric" organization of data collection process.

6.1.4. Organization of user testing and bug repair
User testing and bug repair process on some modules was not organized efficiently enough and it did not enable to complete the user testing and bug repair phase timely.

6.1.5 Some recommendation for the future:
We highly recommend to perform smaller projects with strong project leaderships; even smaller projects need two phases: the first is the preparation of business redesign (could be paper-and-pen, no prototypes yet) and the second is the translation of redesign into IT language. The first phase is time consumming and all the details of the data process should be described and explained.