Cloud for Official Statistics

Progress

The project held a sprint meeting in Belgrade from September 12-14. Leaders or experts of each subgroup theme were present. Other project experts participated online throughout the sprint, some at very early or late hours. Much collaboration, exchanges and progress were made. The sprint concluded with a webinar. Its objectives was to introduce the project, provide an overview on each of its five theme and leave much time for questions, opinions, experiences and other relevant information from the audience. The webinar was attended by 40 persons.

A main draft document was created. Each subgroup has inserted their respective documents in it and are now working from the main draft. Formatting and editing the document has started. The draft document is due to be completed by November 1st.

Next Steps
Complete the draft document on November 1st Deliver a webinar to present the content of the document. It will be delivered on November 16 at 12:00 (CET). An announcement will soon be sent. We hope to have many members of the HLG-MOS community present! Finalize the document in December.

Risk and Issues

Issue

Mitigation

Future of collaboration and sharing on the adoption of cloud,

At this point, we do not see the need to extend the project with a commitment to deliver other documents. However, we plan to conclude the project by determining if there is a need and commitment in continuing to share adoption experiences among statistical organisations. If so, the current project could setup the main parameters of this cloud adoption community (how, who, main topics, frequency, etc.)

The idea of setting up a cloud adoption community will be presented and discussed at the webinar and the HLG-MOS workshop
This community will need more organisations than currently collaborating on the project

ModernStats Carpentries (phase 2 Meta Academy)

Progress

Work Package 1: Following Kate's departure from Statistics Canada, UNECE Andrew Tait has taken over the coordination As part of WP1, the Python curriculum is nearly final (see Python for Official Statistics) and the contributors believe completion is achievable by year end pending final review. The other envisaged curricula (R for Official Statistics, Git for Official Statistics) are less advanced and would require a project extension for being delivered. As part of the closure workshop (see WP2), WP1 outcomes were presented to Carpentries representatives.

Work Package 2: A closure workshop was organised with the Carpentries in October, with the participation of WP2 and WP1 contributors. The workshop was organised around 5 questions formulated in July by the WP2 contributors:

Question 1: How can we incentivise trainers to fully embrace the framework?
Question 2: Is the Carpentries IP model acceptable for public sector statistical organisations?
Question 3: How does the governance of the Data and Software Carpentries work?
Question 4: Are we comfortable becoming direct members of the Carpentries?
Question 5: Do we all agree on the preferred, most realistic approach? Open discussion on next steps in 2024-25.

The summary record from the workshop is being finalised.

Next Steps

The information harvested will be formatted into the final report for the HLG-MOS meeting in November. It appears that the one realistic scenario could be to create an "Official Statistics" curriculum under the Data Carpentry, as an intermediate step before creating a full fledge "ModernStats Carpentry". This could be materialised with a the 3 envisaged curricula (Python, R and Git) going through the Data Carpentry curation procedure (starting with the Lesson Incubation stage).

Going through that process and delivering an "Official Statistics" curriculum under the Data Carpentry, could be the scope for a 2024 project. But that would require an organisation stepping up to take over the lead / coordination by Statistics Canada on WP1 in 2023.

Risk and Issues

Issue

Mitigation

The WP1 Lessons are all still incomplete, with only the Python lesson scheduled for progression to Alpha stage by year’s end. The lessons

Additionally, WP2 topics could use further investigation.

An extension of the project thus makes sense, however with the departure of previous WP1 lead Kate Burnett-Issacs of StatCan, any extension would need to have someone replace her, otherwise the project should be put on hold.

Jonathan Wylie of StatCan has been taking up Kate’s work and is currently bringing himself up to speed with the Carpentries project.

Stéphane Dufour will contact him regarding further collaboration on the project.

Discussion of an extension to the project will occur at the workshop.

Data Governance for Interoperability Framework

Progress
The final document is almost completed: chapters 1 and 2 need only a final editing, chapter 3 will be soon be aligned to 1-2. We have still 2 meetings to finalize chapter 4 on Recommendations and to prepare a version presentable at the workshop

Next Steps

Final editing for chapters 1-3, completion of chapter 4 in time for the workshop

Find a way to discuss the document inside the Community (webinar?)

Verify connections with any future activity on CSDA

Risk and Issues
Issue	Mitigation

News from the Groups
Blue-skies Thinking	Identifying Topics/Opportunities	CONTINIOUS	Potentially we may receive a proposal from Ian O’Sullivan on the Survey Playbook pitch, i.e. a living document that provides advice and guidance on best practice on surveys for NSOs. However, if the proposal is not ready in time, then discussions may continue in other fora such as the next Data Collection Expert Meeting. The Survey Integration pitch received earlier this year will not progress as a proposal/activity at this stage, as Ian is awaiting confirmation of internal support on the subject, and there are no guarantees such will be received before the workshop.
	Non-Probabilistic Surveys	IN PROGRESS	Also known as “Smart Surveys”. No project/activity proposal received at this stage. Further discussion make take place at the HLG-MOS workshop.
	Digital Twins	IN PROGRESS	A presentation was given in September outlining the status of Digital Twins. Current work allows for testing design decisions, with future progress aimed towards a tool for planning and monitoring. No project/activity proposal received at this stage, but discussions will continue within the modernisation group.
	Open Source Adoption	IN PROGRESS	The BSTN group has received a proposal from Barteld for consideration at the upcoming HLG-MOS workshop. The proposal is regarded by the group as a good launching point for discussion, which could lead to some other project proposals, given the large scope of the topic.
	Other - "The use of LLMs for Official Statistics" white paper	Work continues on the draft paper. Sections 2-4 are almost complete, with mainly polishing to be done. However, Sections 1 and 5 still require further development. The deadline for completion of Sections 2-4 is November 3^rd.
Applying Data Science and Modern Methods	Implementing ML-based Solutions in Data Editing	IN PROGRESS	The team has been actively presenting and discussing individual contributions related to selected chapters, as well as finalizing the introduction section. The primary objective for our upcoming meeting is to address the remaining chapters and ensure the document maintains a coherent structure. We are currently in the process of confirming with the use case authors to ensure that the use cases are up-to-date and to secure permission to include them in the final document. Consideration is being given to 'where to next'.
	Understanding and Selecting Models	IN PROGRESS	The team is currently working on two separate documents. The first document provides research and guidance for using "LLM's for methodological advice," while the second document addresses the outcomes of discussions on algorithms and their relationship with existing standards. The team acknowledges a slight delay and aims to have draft documents ready for conclusion in December.
	Framework for Responsible AI	IN PROGRESS	The team made some progress on the three deliverables: Deliverable 1: Guiding Document - All chapters are nearly complete. Our next step is to have these chapters reviewed by both internal and external reviewers. Deliverable 2: Assessment Tool (Checklist) - This will be finalized once the guidelines have been completed. Deliverable 3: Review Process - We have had productive discussions with the team about a draft proposal for the review process. However, it requires some finalization. Consideration is being given to 'where to next'.
	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, with the authors for the various sections delivering their respective sections. The whitepaper is expected to be ready for the November HLG-MOS Geneva meeting. All the updates will be published on the wiki (access limited).
Capabilities and Communication	Work and Job of the Future	IN PROGRESS	The Growth Generic Growth model is ready, it deals with topics such as inclusion, reaching youth, flexible workspace recruitment. etc.
	Data Analytics	IN PROGRESS
	Ethics Management (Data and Business)	IN PROGRESS	Progress on the reference book on ethics - colection of specific examples of the methods that organisations use to foster integrity and ethical behaviors
	Communication	IN PROGRESS	Document describing communication through the lens of inflation crisis and comunnicating CPI is at the final stage.
	Ethics Workshop 2024 (26-28 March 2024)	IN PROGRESS	Invitation to the Workshop on Ethics was successfully sent to the countries. Deadline for the submission of abstracts is end of November. Workshop is planned to be in person in Geneva on 26-28 March 2024. You can see more information here: https://unece.org/info/events/event/383575 Next steps: To form Organizing Committee of the Workshop.
	Other
Supporting Standards	GSIM Revision	IN PROGRESS	A small group completed the review of the feedback provided and it's currently implementing the few changes to the model.
	GSBPM-GAMSO Revision	IN PROGRESS	The task team has most recently examined feedback regarding the "Process" and "Collect" phase. An important decision has been made to change the name of phase from "Collect" to "Acquire". It was suggested to make "Integrate" sub-process as a separate phase, but there wasn't a strong support for such big change. For more details, please check our Github group @ https://github.com/UNECE/GSBPM_GAMSO_Revision/
	SDMX-DDI-GSBPM	IN PROGRESS	Two sub-groups, one specialized on DDI and the other on SDMX, met to complete and clarified the use of both standards in individual GSBPM sub-processes and phases. VTL will also be included wherever appropriate. The plan is to meet again by early November and complete the DDI and SDMX descriptions of the GSBPM core focusing on each sub-process on the standard that provides the best fit.
	Core Ontology for Official Statistics version 2	NOT STARTED	No update.
	Other	Discussion on ModernStats World Workshop (MWW) 2024 continues. At this point we are trying to satisfy the many constraints there are around the Conference On Smart Metadata for Official Statistics (COSMOS) which takes place in Paris in April, and whether having the event is feasible. No decision yet. Discussion of task team proposals for next year started. The team is considering a proposal for a revision of the Common Statistical Data Architecture (CSDA).