Issue #2-2: More details to Business Group

The implementation of the Programme of statistical activities of the Republic of Croatia (2013 – 2022) is achieved by the adoption of the Annual Implementation Plan of statistical activities of the Republic of Croatia (in further text Annual Plan). Annual Plan is adopted for each calendar year. It uses the thematic structure of the Statistical requirements compendium, as well as the Programme. 

All the statistical activities carried out in the Republic of Croatia, which are held by the Croatian Bureau of Statistics and other administrative organizations and institutions, are described in the Annual Plan. Based on Annual Plan, CROMETA (CBS central metadata repository – core of the system for processing statistical activities) has included metadata on statistical activities and supplemented with contact information and other metadata. The CROMETA database model is designed according to the so-called reference model created in the Eurostat METANET project (2000-2003). 

Within the CROMETA model there are different types of metadata according to purpose. Some metadata are used for a declarative description of the content of statistical activities while others are process-oriented and describe ways of collecting and processing data. Along with them there are global concepts and methodologies used independently of statistical activities, as well as general metadata describing organizational structure, authorization and access rights.

Annual Plan is described as a hierarchical structure constructed of thematic units. This hierarchical structure can have a different number of levels throughout the years. Thematic units are stored in database as metadata type named Theme, and the entire hierarchical structure of Annual Plan is stored in the metadata type Theme Structure. Any change in the structure of the upcoming Annual Plan, such as ordering or marking thematic units, entails adding new versions of thematic units (Themes) that will conform to the form prescribed by the current Programme of statistical activities.

Metadata type Study is created for a description of statistical activities that are conducted in CBS. Study is linked to one Theme and covers all encompasses all types of statistical activities (statistical activities based on data collection, statistical activities based on administrative data sources and development activities or projects – these are stored in database as metadata type Activity Family). Also, concept Study serves to unite the versions of statistical activities that are different for years.

In the hierarchical representation of the statistical activity for a given year, actually, there are no Studies, but their versions that were carried out in that particular year. This is another kind of metadata, the Study Version, which saves versions that are primarily different to the research period. The activity (Study) can be the same for many years, but every year a new version of statistical activity (Study Version) is added, the period of which is within the year for which Annual Plan is being generated. If thematic unit (Theme) does not change within Annual Plan structure, then one activity (Study) can have multiple versions (Study Versions). Only one (usually last) has the tag that is valid. 

Metadata on statistical activities are the starting point of navigation in multiple applications developed under CBS’s integrated statistical information system, and present process and survey data. To facilitate the entry of these metadata, CBS also developed a web application from which an Annual Plan is generated, and is only available to authorized persons who add metadata on statistical activities for the following period based on previous years.

We have recognized this issue. It would be important to be able to follow, for example, metadata created in different parts of different processes. This is also an important issue for the renewing of our ERP solutions which are under investigation at the moment here at Statistics Finland. We do not have any experience on a recommended solution, but we would be interested to learn from the other organisations.

We agree with the proposal of adding objects related to Statistical Program, Business Process, and statistical Program Cycle.  However, the point is not defining these concepts, as they refer to processes or executions of GSBPM.  The focus should be the definition of the relevant data contained in the objects. 

We do also have processes that used as input the output of other processes.  For instance, National Accounts uses as input the result of many of our economic statistics.  A second example could be the analysis of satellite images to produce statistical information.  The combination of the data gathered from combined sets of sources included statistical programs is a regular way to generate statistical indicators.

Nevertheless, we see every production line of a program as an independent entity.  We have a full GSBPM cycle for every production of information and we concentrate in the modelling of this line (mapping GSBPM and GSIM).  We have not tried to fit into one model the execution of parallel or serial execution of programs.  The number of combinations will be extremely high; the model would be too complex; and of little general use.  In summary, we recommend not to integrate interactions of cycles.  We might try to develop some white papers where we document how to use GSBPM and GSIM in national accounts.

As per the wiki contribution from the ABS on 10 May 2018, ABS has done a lot of work in this area at an implementation level although it does not necessarily result in changes in the high level conceptual model.

The ABS approach supports scenarios such as that outlined in the above question, and others.

Working through these scenarios has proved complex for reasons such as

1. Some Statistical Programs (SPs) have “ongoing” (rather than cyclic) data acquisition activities, but may produce outputs on a cyclical basis to be consumed by other SPs and/or to be disseminated to the public. A consistent and simple “end to end” cycle is not necessarily associated with each SP.
2. One applied use for SPs in the ABS relates to access control, but sometimes the current implementation can lead to pressure to subdivide SPs for security purposes where from a business perspective they are a single set of high level business activities. This appears to be a solvable issue by updating implementation.

It highlights, however, that a lot of implementation issues arise at hand over points between

• Statistical Support Programs (eg a centralized data acquisition function) and SPs, or
• one SP and a second SP that requires data from the first as an input.

Overall ABS agrees with the wiki contribution from INSEE on 10 July 2018 that analyzing the extent to which changes could be warranted to the conceptual model would be a huge piece of work. It would need to start, in any case, with implementers sharing their approaches to applying this part of GSIM to identify commonality of requirements, approaches and challenges.

Even if such an activity led to a “practice guide” for implementing this part of the model, rather than changes to the conceptual model itself, it could be valuable as a separate activity under the Supporting Standards Group.