6 th SDMX Global Conference 2 5 October
6 th SDMX Global Conference, 2 -5 October, 2017 My reading of the ‘’SDMX Starter Kit For National Statistical Agencies’’ Rafik Mahjoubi Rafik. mahjoubï@oecd. org
Starter Kit the context • 2014 ADB and UNESCAP SDMX initiative in Asia and the Pacific • Based on draft new version 2017 – work in progress • By Denis Ward (teedward@gmail. com)
The Starter Kit chapters A. OBJECTIVES OF STARTER KIT B. BUSINESS CASE FOR IMPLEMENTING SDMX STANDARDS AND GUIDELINES C. STRUCTURED PROCESS FOR SDMX IMPLEMENTATION BY NATIONAL AGENCIES D. DATA STRUCTURE DEFINITIONS AND METADATA STRUCTURE DEFINITIONS E. SDMX CONCEPT SCHEME AND INFORMATION MODEL F. DATA REPORTING, EXCHANGE AND DISSEMINATION SCENARIOS G. BUSINESS PROCESS MODERNISATION - OTHER RELEVANT STANDARDS H. SDMX IMPLEMENTATION TOOLS I. MOST COMPLEX / DIFFICULT ISSUES FOR COUNTRIES NEW TO SDMX TO COME TO GRIPS WITH
A. Objectives of Starter kit • A resource for countries with limited SDMX knowledge • Process for developing a business case • Goes beyond the usual focus on content and technical SDMX aspects • NSDS/ Statistical master plans frameworks • Structured implementation process- including prior SDMX
B. Business Case for Implementing SDMX … Theme SDMX relevance Advocacy Reliance on external resources and need to improve advocacy within national statistical systems (NSSs) Agencies to demonstrate the relevance of their outputs to the major problems / issues facing the country. Coordination Statistical organisations reported the need to improve coordination + absence of guidelines for setting national standards across the statistical collection / compilation / dissemination cycle. Whilst SDMX can provide the technical tools to alleviate these problems, there is a need for NSS to first reach agreement on key issues such as responsibilities, organisation of data flows and systems architecture… Infrastructure All NSOs report the need to improve statistical infrastructures including registers, metadata repositories, databases, data dissemination tools and facilities. Statistical integration necessitates the use of standard concepts and code lists and the elimination of domain-specific systems, etc. , Skills All NSOs highlighted the importance of developing the skill capacity of staff Related to capacity development is the need to develop strategies for the retention of key skilled staff in order to maintain an adequate core of required expertise. If high staff turnover cannot be avoided then strategies / processes need to be developed to minimise its impact.
B. Business Case for Implementing SDMX … Problems faced • Data are collected in various ways (surveys, web queries, data files, metadata text, emails, etc. ). • Data collection is a labour intensive, manual operation. • Possibility of errors and inconsistencies. • Turn statistics production into an ‘’Industry’’
B. Business Case for Implementing SDMX … How SDMX can help • A common Language and model • Facilitates the development of standard tools • Industry standard allows separate implementations interoperate output and input • SDMX reduces reporting burden • SDMX help standardise IT applications and infrastructure • Etc.
C. Structured Process for SDMX Implementation by National Agencies
C. Structured Process for SDMX Implementation by National Agencies 5 Steps STEP I: Acquire a basic understanding of key SDMX artefacts 1. SDMX Standards and Guidelines 2. SDMX Registries 3. The importance of using common and harmonised language and concepts STEP II: Addressing issues to be considered by countries prior to embarking on SDMX implementation 1. Institutional issues and objectives 2. Statistical issues 3. IT-Related issues 4. Skills-Related issues 5. Preparation for a small-scale pilot SDMX implementation a. Practical tips b. Issues to be considered when preparing for a pilot project STEP III: Identification of national SDMX capacity development needs and capacity development delivery 1. Develop and understanding of the SDMX information model 2. Develop the capacity to access, install and utilise SDMX implementation tools 3. Capacity development delivery 4. Internal change processes STEP IV: Determining which set of SDMX implementation tools to use Downloading and use of selected SDMX implementation tools STEPV: Participation in SDMX networks
STEP I: Acquire a Basic Understanding of Key SDMX Artefacts • Review of the literature listed in the starter kit: – sdmx. org – Content oriented guidelines – User guide – SDMX Tutorials by Eurostat – SDMX road map 2020 • Implement a pilot • Use Global or Shared DSD
STEP II: Addressing issues… prior to embarking on SDMX implementation
STEP II: Addressing issues… prior to embarking on SDMX implementation Institutional issues and objectives: • Centralized vs decentralized systems • Firmly link SDMX to NSDS • Collect and/or disseminate and report • Decide on the main drivers in the agencies; the data domains; the data flows… • Establish effective coordination in the NSS, negotiate Institutional, Statistical, IT and resources
STEP II: Addressing issues… prior to embarking on SDMX implementation Statistical issues: • Conformity to international standards: concepts / definitions, structural metadata and related codes… • Identification of the appropriate DSD • Metadata-related issues – very important!!! • Use of a standard metadata template (SDDS of IMF, Eurostat’s Single Integrated Metadata Structure etc. )
STEP II: Addressing issues… prior to embarking on SDMX implementation IT issues:
STEP II: Addressing issues… prior to embarking on SDMX implementation Skills-Related issues: • Financial, staff resource and capacity and skill development requirements • Resources will only be forthcoming if a proper business case is developed • Development of strategies for the retention of key skilled staff- or to limit its impact…
STEP II: Addressing issues… prior to embarking on SDMX implementation Preparation for a small-scale pilot SDMX implementation: • Never underestimate the time and resources • Number of internal and external parts • The domains to include • Nature of the SDMX exchange process • Outcomes firmly linked to full scale implementation • • Better estimate efficiency that may be achieved Accurate estimates of the resources needed Determine the most appropriate means of implementing full scale Experience gained in negotiating and coordinating with one partner will be applied to a large group in full scale
STEP II: Addressing issues… prior to embarking on SDMX implementation Preparation for a small-scale pilot SDMX implementation: • The statistical domain and the DSD • For what: • • Data and metadata reporting to international Org. Efficiency exchange among agencies within the country Agg. data and metadata dissemination to public Improve storage of structural and reference metadata? • Which unit will drive? • The cost?
STEP III: National SDMX capacity development needs and delivery Capacity development needs on: SDMX Tools Statistical SDMX-IM Institutional Metadata IT Internal Change processes Capacity development Delivery
STEP III: Determining which set of SDMX implementation tools to use • • No comparison of tools available Different tools for different needs Operational tools Reliable custodian agency - sustainability Cost and adherence to agency policies Documentation Trainings courses and online tutorials
STEP III: Participation in SDMX Networks Collaboration with advanced systems Join SWG and TWG Expert meetings Global conferences Links to regional and sub-regional organisations • Discussion forums – linkedin: SDMX group • • •
I- Most Complex/ difficult issues… Issue Suggestions Getting senior management buy-in – Develop a Business Case for pilot or full-scale SDMX provision of resources implementation Matching national concepts and code lists Provide examples of national tables / datasets with SDMX specific and international expressed in terms of the SDMX Concept Scheme. terminology Coming to grips with the SDMX Information Provide examples of national tables / datasets Model (SDMX-IM) expressed in terms of the SDMX Concept Scheme. Which Data Structure Definition (DSD) to Refer to SDMX discussion paper, Modelling a use; how many DSDs to use; unmatched Statistical Domain for Data Exchange in SDMX dimensions / code lists. Which SDMX implementation tool(s) to use Refer to comparison information on a range of existing implementation tools provided in Starter Kit
I- Most Complex/ difficult issues… Issue Suggestions Placing data / metadata exchange Refer information in the Starter Kit on how to assess scenarios in context of current and relationship between current / future national IT proposed future national IT platforms platform(s) where data / metadata reside and various exchange scenarios. Relationship between SDMX standards Refer discussion in Starter Kit on the other main relevant and guidelines and other standards, in standards particular, DDI Identifying pilot project modalities Skill / capacity development Refer pilot project tips and tricks provided in the Starter Kit Refer discussion on Business Case, skill needs assessment / capacity development delivery and participation in SDMX networks
Thank you! rafik. mahjoubi@oecd. rog
- Slides: 23