Examples of Implementing GSIM
57. GSIM v1.0 is the first public release of GSIM. As such, statistical organizations have not yet had the opportunity to implement it. However, a number of organizations have started to plan how they will use GSIM.
58. This section contains short statements from some of these organizations on how they plan to use GSIM. At the moment, these statements represent intentions, not results. As GSIM is used by more organizations, there will be further examples of its use.
59. Annex B provides a template which statistical organizations are encouraged to complete regarding local information models and the use of GSIM.
60.Developing GSIM has been a large, collaborative effort from agencies across the globe. The aim was to deliver a product that could be used to benefit the international statistical community, as well as individual agencies, as we seek to modernize the production of official statistics.
61.Statistics New Zealand has identified three initial uses for GSIM. These do not depend on the shared adoption of GSIM. They represent the value that GSIM offers individual agencies. Using GSIM allows an agency to benefit from the collective expertise of the international community, to ensure that terminology, processes, and systems design fit with best practice. In future there will be collective benefits, but in the meantime GSIM can be put to practical use to help Statistics New Zealand transform the way it delivers official statistics.
To collaborate within the organization and beyond
62.GSIM will inform our terminology as we work towards standardisation across the organization. Using the same language will help our teams, including IT and statistical professionals, to collaborate better. We will use GSIM to help new staff understand the basic concepts in the production of official statistics. This will help build our staff capability in understanding the information needs of the statistical production process, and ensure they can more easily rotate around roles in the organization.
63.Using a common model and terminology between agencies will be a strong foundation for international collaboration. The time currently required for projects to get up and running hampers the collaboration effort and reduces the time we spend working towards the project goals. A common model will reduce any miscommunication and misunderstanding.
64.Reducing the barriers to working with other agencies will ensure we better align our thinking with others, and allow us to consolidate our joint vision for the future of statistics.
To underpin and understand existing standards
65.GSIM provides a complete picture of the information that underpins the statistical production process. Existing international standards provide only a partial picture and do not meet the specific needs of statistical agencies. Use of standards such as DDI, SDMX and others such as geospatial standards will be enhanced when looking through the lens of GSIM. We will use the model to support our understanding and help us to determine where these standards meet our needs, where gaps exist, and where future development should be focused.
66.As a reference model, GSIM will help to determine how existing standards complement each other. For example, contributing to analysis of where needs can be best met by using DDI, areas where needs can be best met by use of SDMX, and areas that can be best met by using other standards.
67.GSIM will support the need to develop guidelines and standards that are independent of specific implementation standards. The terminology and definitions in GSIM are meaningful and familiar to statisticians, and a single model covering the entire statistical production process will enable a more consistent overall view of our data and information.
To inform system design
68.We will use GSIM to inform information model designs behind our systems and as a benchmark and starting point for design. It will help provide assurance that our own systems are developed to a high standard.
69.Using the model will improve the efficiency of our development projects by avoiding re-analysing our information needs for each and every project. For example, we are already using the model to inform our development of metadata management systems. As we develop our new classification management system we will aim to be aligned with GSIM.
70.With over 100 projects currently underway to transform the organization, GSIM has a huge potential to improve our processes. Using GSIM during the development of several new systems will be an opportunity to identify successes and areas for improvement.
71.GSIM will help us identify our most high-value and critical data so we can appropriately resource them.
To lay the foundations for the future
72.Many of the initial uses for GSIM we're planning are based on the assumption that at present the model is not yet mature enough, and without sufficient detail, to implement into systems. GSIM is a conceptual model.
73.We hope that as GSIM is developed in the coming years, the information model will be able to be directly implemented and systems built upon it. GSIM will begin to be implemented into our statistical infrastructure systems to enable standardized information exchange across Statistics New Zealand.
74.In the future we will be able to reduce the time it takes to develop new systems – the base information model will already exist – and where other agencies have developed systems based on the same model, these will be able to be repurposed for our own needs.
75.We hope to build our capability by enabling our use of solutions from other statistical agencies and contribute to enable others to benefit from the flexibility of our own new solutions.
76. The Australian Bureau of Statistics (ABS) considers GSIM to be a key enabler of practical collaboration and sharing across a number of levels. This includes at the international level, and also for guiding and facilitating organization programs such as the ABS2017 transformation program.
77. Announced in February 2010, the ABS2017 program aims to transform the way the ABS collects, collates, manages, uses, reuses and disseminates statistical information. This will improve the usability, value and timeliness of statistics for government and the community.
The approach to adopting and applying GSIM as a common reference framework across the ABS is expected to have many analogies with the approach used with GSBPM GSBPM is used as part of the ABS Enterprise Architecture (EA). It provides a common point of reference when planning, specifying, managing and coordinating individual projects and initiatives across the ABS. GSBPM is used as a common set of terminology and common basis for categorization of statistical production activities..
78. GSIM includes a level of detail (the Specification Layer) which is not present in GSBPM. This will support a number of uses of GSIM such as the ability to map GSIM to existing implementation standards like DDI and SDMX.
79. It is expected that GSIM will continue to evolve. ABS will take into account the fact GSIM isn't yet as mature as GSBPM. It will take into consideration the extension guidelines defined in GSIM and determine the impact of GSIM on ABS business requirements.
80. A small task team has been established within ABS 2017 to plan and co-ordinate adoption of GSIM in the first half of 2013.
81. The task team will formalize the inclusion of GSIM as a reference framework in the ABS Enterprise Architecture. GSIM will be a core artifact within Information Architecture but also influence the description of Business and Applications Architecture.
Using GSIM at a business level
82. The task team will gather feedback from various ABS initiatives which seek to apply GSIM as a reference framework.
83. One such initiative is currently reviewing the diversity of terms and concepts currently used by subject matter statisticians across the ABS to refer to Statistical Metadata. The project aims to harmonize the terminology used within the ABS in future.
84. Wherever possible the reference terms to be used in future will be drawn directly from GSIM. Where there is an agreed ABS business need to diverge, the relationship between ABS usage and GSIM usage will be documented. Some, but perhaps not all, of these divergences may result in a proposal to update GSIM terms and/or definitions in the next version.
Implementing GSIM in the ABS
85. GSIM is also interlinked with the ABS' development of metadata infrastructure, in particular the Metadata Registry and Repository (MRR) and the ABS Transitional Metadata Model (ATMM).
86. The Metadata Registry and Repository is a core part of the ABS infrastructure, and will make capturing and reusing metadata much easier and more efficient. The MRR is made up of two parts:
- a repository that will store statistical information (including metadata, data, process definitions, etc).
- a registry or catalogue that will allow us to register metadata, easily search for that metadata, and discover and retrieve metadata held in the 'store' part of the MRR.
87. The ATMM is the model that underlies the MRR. It defines the way statistical information is registered for discovery. The model combines the influences of a number of international standards (including GSIM) with the core requirements of ABS processes and gives a core set of information objects for use across the ABS.
88. The ATMM could be described as the 'operationalization' of GSIM (or the ABS specific version of GSIM). It has enabled the ABS to bring forward the development of the MRR, and to provide valuable input into the GSIM development process – experts working on the ATMM and MRR projects have contributed to GSIM development.
89. As GSIM itself has become more fully developed over time (particularly from v0.8 onwards), it has begun to have a greater impact on the continuing development of the ATMM, further bolstered by the learnings and experiences of the team members who have directly worked on GSIM.
90. Using GSIM together with other implementation standards to create the ATMM and MRR is just one example of local implementation, but it is proving to be an important one.
91. The IMF Statistics Department intends to use GSIM (and GSBPM) at the point in time when we are looking to upgrade/re-engineer existing processes in the context of the Department's on-going "Streamline, Standardize, Automate" initiative. These opportunities will be used to build a holistic view of the end-to-end statistical process.
92. The first activity will identify the sub-processes and information objects used by the IMF in specific instances of the statistical process.
- Create an "as is" inventory of sub-processes.
- Identify core characteristics of each sub-process, such as description and owner.
- For each sub-process, identify its inputs and outputs.
- Match each IMF sub-process to a sub-process in the GSBPM.
- Match each input and output to an information object in the GSIM.
93. The second activity will utilize GSBPM and GSIM to look for opportunities to improve the effectiveness and efficiency of the IMF's work. Analysis that we may undertake to identify such opportunities may include:
- Studying organizational distribution of activities where IMF sub-processes occur in different work areas, but relate to the same GSBPM process.
- Investigating situations where the same GSIM information objects are used as inputs and outputs to different IMF sub-processes, to identify potential areas for standardization and automation.
- Exploring the extent to which existing standards and technologies (e.g. SDMX) can be leveraged to provide the basis for such standardization and automation.
94. These two activities will be conducted iteratively, starting off with a pilot of the sub-processes being worked on as part of an existing automation project, with the aim of organically spreading to a wider range of sub-processes as further automation projects are initiated. This will allow the "bottom-up" description of existing work processes and information objects (the "as is" inventory) to be analyzed within the framework of the "top-down" generic models. At the same time, the understanding and use of the generic business process and information models can be gradually widened and strengthened across the Department.