116. The eight phases and the sub-processes within each phase of the GSBPM provide a set of building blocks that can be assembled in a sequence to create a production process. Section 2 describes the activities related to geospatial information and services that can be carried out in each phase to produce high-quality geospatially enabled statistics following the principles of the GSGF.

117. Some activities, however, are not limited to a certain phase of the process but rather apply throughout the entire production process (e.g. quality). Also, there are activities that should be conducted at a corporate level rather than as a part of specific production process as they support and influence many production processes across the organisation (e.g. management of address register, establishment of Memorandums of Understanding with other government agencies). In the GSBPM, activities of cross-cutting (across the production process) nature are modelled as “overarching process” 1  while activities at a corporate-level are covered by the Generic Activity Model for Statistical Organisation (GAMSO) 2 , another HLG-MOS model complementing the GSBPM. Section 3 describes activities that are of cross-cutting nature and/or corporate-level.

Strategic collaboration and cooperation

118. Geospatial information are fundamental national information assets that can be used as a basis of numerous civic and commercial activities. A geospatial data ecosystem consists of various actors from both public (e.g. NGIA, transport department, agriculture agency, space agency) and private sector (e.g. utility companies, GIS service provider) providing and using a multitude of geospatial information and services. Therefore, coordination and cooperation within the ecosystem are critical to maximise synergy among different actors and avoid duplication of efforts. Depending on the regional and national context, coordination mechanism may vary (e.g. the INSPIRE directive of the European Union, the Sustainable Development Goal (SDG) data governance board in Ireland) 3 . Regardless of the format, active participation of statistical organisations in national geospatial information governance mechanisms, and continuous engagement with other agencies are important to ensure that standards, models and technologies are aligned as far as possible across the geospatial data ecosystem. This participation and engagement with the broader community also helps to ensure that the needs of statistical organisations are communicated and addressed in an efficient manner.

119. Geospatially enabled statistics with a high spatial resolution can provide invaluable input to other government agencies, in particular, to quickly respond to disasters and crises. Statistical organisations can play an important role in the geospatial data ecosystem not only as a producer of various statistical and geospatial information, but also as a provider of data integration services based on its extensive technical expertise. This service is becoming more and more essential in solving multi-faceted issues of the society such as climate change and migration. Statistical organisations can forge strategic collaborations with other agencies to produce high quality data integration products, which can demonstrate the vital function that the organisations can play in evidence-based decision-making processes with increasingly complicated challenges.

Metadata management

120. Metadata provides essential information to understand and interpret the data (e.g. structure, classification used, analysis methodology, quality). It plays a key role to facilitate sharing, querying and discovery of data and services in an increasingly vast pool of data assets in statistical organisations. The use of metadata is not only limited to data products or services, but also includes various information that influences, triggers and regulates production processes (i.e. metadata-driven processes). The importance of metadata and its management is widely accepted in statistical organisations and much work has been done to develop standards and systems to manage metadata associated with typical statistical production process at the corporate level.

121. Compared to this statistical metadata, there is limited awareness and understanding in statistical organisations on the metadata associated with geospatial information and services. Geospatial information include various types ranging from orthoimagery (e.g. satellite data), elevation / depth, water to transport network 4 , and this great variety of structures / formats as well as methodologies / technologies involved in producing the data adds challenges for statistical organisations to standardise geospatial metadata and systematically manage it.

122. Given the wide scope of geospatial metadata, it is important to first investigate crucial metadata elements (e.g. data type (point, line or polygon), time stamp, coordinate system) needed for different stages of production and determine a core metadata set and standards (e.g. ISO 19115, ISO 19119, GeoDCAT) to follow at the corporate level. After priorities are defined, a continuous improvement process could be put in place to gradually improve the scope covered by the metadata. It is important to have a corporate strategy in place to build a consistent metadata system to avoid compliance issues with existing metadata systems. As for statistical metadata, the geospatial metadata should be managed and updated continuously throughout the production process as the changes affect downstream tasks and influence the final outputs.

123. Alignment and harmonisation of geospatial metadata concepts with those of statistical metadata in existing metadata systems is critical, and there has been an increasing effort to connect statistical metadata with geospatial metadata (e.g. technical specification expanded for geospatial metadata in SDMX 3.0 5 ). Emerging semantic web standards (e.g. Resource Description Framework (RDF) vocabularies) can provide the flexibility in modelling metadata as well as data 6  and its use for disseminating on the web offers a great potential to link outputs of statistical organisations with the plethora of data and resources on the web. Statistical organisations are encouraged to explore the semantic web standards as a long-term strategic objective with successive milestones to achieve dissemination of data and metadata within the framework of Linked Open Data (LOD).

Quality management

124. Quality is one of fundamental characteristics that define official statistics and its management throughout the production process has been a critical issue for statistical organisations. Quality is usually defined in terms of several dimensions on which various quality metrics are developed and agreed at the corporate level to ensure that quality is documented and monitored in a consistent and systematic manner for different processes across the organisation.

125. With growing use of administrative sources and the shift from single process / product to multiple processes / products, however, the management of quality has become increasingly challenging for statistical organisations. Examples of complications related to production of geospatially enabled statistics are grouped by input, processing and output aspects as below: 

  • (Input-aspect) For some types of geospatial information (e.g. earth observation data, network data), understanding the data quality often requires technical knowledge of the field 7 ;

  • (Input-aspect) As statistical organisations move toward high-resolution products and point-based geocoding, the quality of geographies used in the production as an input has a greater impact on the process and the output, compared to when geographic units were at a more coarser level (e.g. provincial, regional);

  • (Input-aspect) While geospatial information may have quality terminologies similar to those in the field of statistics, they might be interpreted and calculated in a different way (e.g. term “precision” means spatial resolution rather than statistical variability 8 );

  • (Processing-aspect) Geospatial information is often used as a basis for integrating data from different sources (e.g. survey, administrative data, big data). The quality of this integration is affected by the quality of geospatial information in each input dataset, but also greatly influenced by the quality of the geocoding or address-matching process itself;
  • (Processing-aspect) More thorough disclosure control processing is needed as high-resolution geospatial information carries a greater risk of privacy infringement, not only on its own but also in combination with other data products;
  • (Processing-aspect) There is a lack of understanding of the impact of geospatial information quality across different types of processing in the production process;
  • (Output-aspect) Ensuring accessibility and usability of geospatial information and services could be challenging as there is a wide range of different requirements, priorities and needs depending on the user group. For example, researchers would require microdata equipped with geocodes readily available to integrate with other data sources for their analysis, city and municipal authorities would want datasets to be provided in formats that can be easily integrated within their local system, and journalists would be interested in key information and digestible headlines from geospatial analysis;
  • (Output-aspect) With a widespread use of web-based map services, there is a greater expectation on the quality of map products from statistical organisations (e.g. visualisation, interactivity, user-friendliness).

 126. Quality management that can be conducted as an overarching process or at a corporate level includes:

  • Identifying the authoritative (external or internal) sources of reference data, establishing the quality profile of each data source based on the primary use cases within the organisation and communicating continuously with data providers regarding the quality requirements and associated risks;
  • Establishing mechanisms with which feedback from production instances can be incorporated into the quality management processes of geospatial information holdings such as address registers (e.g. verification of geospatial information during field survey);
  • Developing quality dimensions and metrics to be used at different stages of production and a consistent strategy at corporate level;
  • Monitoring the new developments in the fast-evolving geospatial field and discussions at the regional and global levels to ensure that knowledge, methods and technologies in the organisations are up to date and in line with those of other communities.


  1. For more details about “overarching process”, see GSBPM Section 2 and 4 ↩
  2. For more information about GAMSO, see UNECE Statistics Wiki (https://statswiki.unece.org/display/GAMSO) ↩
  3. For more, see UN-GGIM “IGIF Strategic Pathway 1: Governance and Institutions” (https://ggim.un.org/IGIF/part2.cshtml)  ↩
  4. For more, see UN-GGIM “Global Fundamental Geospatial Data Themes” (http://ggim.un.org/documents/E-C20-2018-7-Add_1-Global-fundamental-geospatial-data-themes.pdf). ↩
  5. Acknowledging the importance of geo-referencing statistical information, version 3.0 of SDMX will include technical specifications to improve the management of associated metadata. These technical specifications will help to connect the different levels of statistical information to geographical characteristics, making possible to include detailed geospatial structural and reference metadata in the exchange of statistical information. This work was done in collaboration with UN-GGIM, which is developing a Geospatial Roadmap to provide a bridge between statistical and geographical information. ↩
  6. RDF-based ontologies such as Web Ontology Language (OWL) and Simple Knowledge Organization System (SKOS) can link resources using pre-defined properties such as owl:sameAs or skos: exactMatch. ↩
  7. For example, see the comparison of quality concepts between statistical community and Earth Observation community in UN Global Working Group on Big Data - Earth Observation Data Task Team “UN Handbook on Satellite Data” (https://unstats.un.org/bigdata/task-teams/earth-observation/UNGWG_Satellite_Task_Team_Report_WhiteCover.pdf). ↩
  8. H. Veregin (1999) “Data Quality Parameters” ↩

  • No labels