Meeting notes - May 9th, 2023

[InKyungChoi]

Participants: Zoltan Vereczkei, Antti Santaharju, Andrea Petres, Carlo Vaccari, Kevin McCormack, Michelle Moss, Gabriel Gamez and Inkyung Choi

Issue #3: GSBPM Phase 8 Evaluate & Quality – Summary: to keep the phase, but improve descriptions so that the differences between phase, OP and GAMSO activities are clear. Also, consider changing the name of phase

Issue #9: Identification of phases and sub-processes: to use numbers or not? – update: as the Core Ontology for Official Statistics (COOS) team is already using GSBPM numbering as part of their URI system, we contacted to ask about impact and how they think about this change.

Issue #4 GSBPM Phase 7 Disseminate

Similar to our comments regarding GAMSO, “Disseminate” is very product focussed, whereas we are seeing much more data delivery so that others (often Private companies) create & disseminate products (effectively move the NSO from a Retailer to a Wholesaler). Such a transition, or the need to support such initiatives, could be better reflected in ‘Disseminate’. (New Zealand)

• It is not clear what this means, in GSBPM, do we need to care whether our users are intermediaries or not?
• Or is it just about wording, for example, moving away from "dissemination" (just putting our products out) to "delivery" (making our products in a more consumable and interoperable way for diverse users)?
• Linked open data has become much more important these days (compared to 2019 when the previous GSBPM came out), it would be worth highlighting it. We might need to do the same in Design phase (because it needs to be designed first), but for now, put everywhere needed (and then remove if there are too many duplicates)
• Action: InKyung to check with Gary; Zoltan to think about potential wording changes

Phase 7. Disseminate is now a mixture of dissemination and communication elements. We suggest removing communication from this phase. (Hungary)
For Phase 7. it is difficult to draw the line between dissemination elements directly linked to the process and general supporting services (example: user consultations). If there is a UNECE guideline available on dissemination, its recommendations should be taken into account during the revision of Phase 7. (Hungary)

• Statistical organisations have been quite bad with communication, this is a big issue
• From the modelling perspective, activities (e.g., communication) can be in any model, shouldn't GSBPM focus more on how we put the activities together
• The phase name "Disseminate" might be legacy, if our ultimate purpose is to satisfy user needs, just putting products out is not enough, we need communication. There have been a lot of developments that make dissemination more automatic/systematic (e.g., SDMX), but communication seems remain more "manual"
• It is important to distinguish communication as a corporate support (e.g., communicating value of official statistics) and communication as a part of production process (e.g., briefing on Census results)

GSBPM does not mention the archiving of finalished statistical outputs and processed microdata on which they are based. Archiving is mentioned only in the context of the metadata management (archiving rules) and in the collect phase. In Finland, the Data Management Act requires the archiving of these finalished outputs and microdata. Statistics Finland considers the formatting of this information to be part of a sub-process 7.1 and updating it in the archive to sub-process 7.3. (Finland)

• The cost of archiving is getting lower while the benefit of keeping track of everything and being able to give access to archived data to specific user groups (e.g., researchers) is getting higher.
• In some countries, archiving at the time of final product release is a milestone activity, but archiving practice varies.
• Archiving should be a part of overarching data management activity, it happens everywhere.
• Preliminary conclusion: not to include in Phase, but improve data management description and use this as a milestone example

Phase 7 Introduction, the suggestion to include cataloging and tagging"...

• Conclusion: agree to add

Clarify the difference between "statistical outputs" and "dissemination products". This will remove the ambiguity between sub-processes 6.5 and 7.2. (France)

• "Outputs" can be produced at any point of the production process, and "product" is a package for dissemination which is meant to reach users.
• Conclusion: to add a footnote explaining the difference in 7.2 (when term "product" starts appearing)

[InKyungChoi]

Feedback from Gary (NZ; May 22, 2023): In terms of the ‘Disseminate’ comment, I’d suggest the wording saying something like “ ….. there are more diverse user groups and that our data & products should be more interoperable and consumable ….”. While it can be argued that data is a product, I think the separation of ‘data’ and ‘products’ strengthens the statement and pushes the reader to consider just releasing data for others to consume, repackage etc.