Cloud for Official Statistics

Progress

Some project subgroups have met and further developed their working documents. The most recent version of the five work documents is posted on the project's home wiki page. A meeting took place with SORS to prepare the sprint that will be taking place in Belgrade from September 12 to 14. The sprint will be attended by 18 project experts (12 in-person and 6 virtually; 3 non-response still). Accommodation options and other key information were distributed to all project experts. The sprint will include a webinar to introduce the project to a wider audience and seek their input. The webinar was announced on LinkedIn.

The bad news is that the project has lost two subgroup leaders (StatCan and ONS). A replacement was quickly identified (Ian Bale) who also volunteered to take the lead of a subgroup. The expert from the ONS is looking into finding a replacement. Both experts left for career opportunities in other government departments highlighting quite concretely the challenges of retaining people with a high level of Cloud expertise and experience.

Next Steps

Meetings to prepare for the sprint will be organized this week and next week.

The sprint will be held from September 12 to 14. Everything possible will be done to optimize the participation and collaboration of all.

The sprint will conclude with a discussion on possible extension. The outcome of this discussion will be communicated to the BSTN, hopefully on September 15.

Risk and Issues

Issue

Mitigation

Loss of expertise. Given that we are late in the year, there is not much we can do except hoping for replacements and integrating them as quickly as possible.

Lower attendance at the sprint. We appreciate all the support that the HLG-MOS committee and its Exec Board can provide to encourage and allow experts to attend.

ModernStats Carpentries (phase 2 Meta Academy)

Progress

WorkPackage 1: Work on WP1 of the Carpentries project is progressing but has been impacted by recent events. The departure of Kate Burnett-Isaacs from Statistics Canada means that others will need fill the gap. Currently there are three GitHub repositories and corresponding GitHub Page sites (on R, Python, and Git) under development, however all are at the pre-alpha stage and at various levels of completeness. One complication is that these lessons will need to be transitioned from their current format to a new Carpentries lesson format (tools to automate the transition process can be found here). Of the three lessons, while each has an overall structure in place, the Python lesson is by far the most advanced. Consequently, discussions have revolved around finalizing the Python lesson.

Work Package 2: after several months of experimenting the Carpentries governance, WP2 was kicked off in July - see presentation and meeting notes here: WP2 - Kick-off. OECD reported on its experience with silver membership and certification of trainers. A discussion ensued on the governance, intellectual property and business model. Key questions were scoped that will be discussed with the Carpentries in a workshop planned end Sept. The proposed scenario for implementation could be (to be confirmed at the workshop):

"continue to experiment for 1-2 years under Governance Scenario 1(under Data Carpentry and Software Carpentry), which requires all participating organisations to become direct Carpentries members to have a pool of a few certified trainers per organisations (Cost 1), who develop lessons under existing Carpentries (Cost 2). For that to happen, we need to have satisfactory answers to all previous questions. Under the HLG-MOS, the HRMT (Human Resources Management and Training) group could naturally act as the coordination level, reporting to the HLG-MOS Executive Board. With this approach, the heavy burden of initiating a separate curriculum “ModernStats Carpentry” is pushed back until sufficient experience is gained by the community." (see meeting notes)

Next Steps

Work Package 1: set up a meeting with the Python lesson lead, consolidate comments to the lesson, and transition to new format.

Work Package 2: organisation of a workshop with Carpentries representative towards the end of Sept, to review all key questions raised in the area of IP, governance and business model. Based on their answers, the preferred scenario will be fleshed out, and presented as part of the report to HLG-MOS meeting in Nov.

Risk and Issues
Issue	Mitigation
WP1 manager Kate Burnett-Isaac has left Statistics Canada	UNECE Andrew J. Tait to take over the coordination role. WP1 delivery to focus on the most advanced part of the curriculum (Python) and put on hold work on Git and R.

Data Governance for Interoperability Framework

Progress

In July hybrid in Ottawa (ISI)

Meetings even during the holiday period

Advancements for the document's Chapters 1, 2 and 3

Next Steps

Complete revision for chapters 1 and 2

Complete first draft for chapter 3

Detail structure of chapter 4 and start drafting

Risk and Issues
Issue	Mitigation

News from the Groups
Blue-skies Thinking	Identifying Topics/Opportunities	CONTINIOUS	There has been several pitches around data collection; it needs to be decided how to proceed with them. The topic on co-development of open-source solutions (e.g., a solution for rapid survey) was discussed. The approach will be about starting from bottom-up to see what kind of practical issues arising when we try to work together and exploring what it means to maintain open source together as a community.
	Non-Probabilistic Surveys	IN PROGRESS	No update
	Digital Twins	IN PROGRESS	Statistics New Zealand made a pitch on the use of digital twins for survey data collection operations. This will lead to a follow-up discussion with a broader audience interested in the topic to understand whether there are opportunities for collaboration. This open online session is scheduled for 19th September (13:00 CET).
	Open Source Adoption	IN PROGRESS	The team needs to identify a new lead (Kate used to be the lead) and decide how to proceed. The topic is gaining a great interest in the official statistics community. There might be a formal expert group to be created under ESSNet; coordination would be needed.
	Other - "The use of LLMs for Official Statistics" white paper	Good progress is being made with the "The use of LLMs for Official Statistics" whitepaper. Starting with a general kick-off meeting, authors for the various sections have been identified, sections have been shaped, and have already concluded three section meetings (with two more to come). All the updates will be published on the wiki (access limited).
Applying Data Science and Modern Methods	Implementing ML-based Solutions in Data Editing	IN PROGRESS	The data editing Task Team has not met since the last report due to the chair being at ISI, and other absences. Some work has begun on drafting sections of our guidelines. The Task Team expect to resume meetings in September.
	Understanding and Selecting Models	IN PROGRESS	No update as the Task Team Leader is on vacation.
	Framework for Responsible AI	IN PROGRESS	The August meeting was canceled, but the team made good progress this summer on the guiding document (deliverable 1): - The team began working on the core guiding principles "Validity and Robustness" and "Privacy and Security." - The team also made strides in drafting guidelines for the remaining core guiding principles. - The first draft of the review process (deliverable 3) has been shared with the team for their input.
	Other - "The use of LLMs for Official Statistics" white paper	Good progress is being made with the "The use of LLMs for Official Statistics" whitepaper, starting with a general kick-off meeting, authors for the various sections have been identified, sections have been shaped, and have already concluded three section meetings (with two more to come). All the updates will be published on the wiki (access limited).
Capabilities and Communication	Work and Job of the Future	IN PROGRESS	Work is progressing well. In September, colleague from the Communication department of Statistics Netherlands will work on the growth models prepared by the team, to make them more presentable.
	Data Analytics	IN PROGRESS	No update
	Ethics Management (Data and Business)	IN PROGRESS	Work on the reference book on ethics is progressing. Team is collecting areas of interest, examples of ethical dilemmas and replicable practices on ethics.
	Communication	IN PROGRESS	Work on the document describing communication through the lens of inflation is progressing well,
	Ethics Workshop 2024 (26-28 March 2024)	IN PROGRESS	Draft information was prepared and will be discussed with the team at 5 September meeting. Invitation letter will be sent to the countries by the end of September.
	Other
Supporting Standards	GSIM Revision	IN PROGRESS	The feedback period has concluded. Feedback will be reviewed and reflected to the final update for GSIM 2.0 and the user guide.
	GSBPM-GAMSO Revision	IN PROGRESS	The task team has most recently examined feedback regarding the "Analyse" phase. The need for additional privacy protection for geospatial data and referential metadata for new analysis techniques such as ML/AI were added. Team started working on "Integrate: phase. For more details, please check our Github group @ https://github.com/UNECE/GSBPM_GAMSO_Revision/
	SDMX-DDI-GSBPM	IN PROGRESS	After going through many GSBPM sub-processes, it became clear that the extent that SDMX and DDI could contribute depends on the nature of sub-processes. The task team will go through sub-processes to identify most relevant sub-processes and focus on them.
	Core Ontology for Official Statistics version 2	NOT STARTED	No update
	Other	The discussion on the ModernStats World Workshop (MWW) 2024 has started. As Conference On Smart Metadata for Official Statistics (COSMOS) takes place in Paris in April, it would be ideal to have MWW2024 back to back with it.