The OECD statistical information system integration of energy data with other domains November 2007 lars.thygesen@oecd.org
Statistics in OECD Statistical information on society many sectors: economy, labour, health, education, governance... tool for policy analysis compare countries identify good outcomes and less good e.g. country economic reviews e.g. education authority & quality November 2007 2
What is the Statistical Information System? Immense integrated statistical information bank Data and metadata can be combined and used over and over again, for innumerable outputs Horizontal products covering various subject matters can be developed at a reasonable cost Provided data and metadata are properly organised November, 2007 3
A Statistical Information System Ques- Tionnaire Ques- Tionnaire Ques- Tionnaire Input1 Input2 InputX Mid1 Mid2 MidX Output1 Output2 OutputX @ 1 @ 2 @ x November 2007 4
Ques- Tionnaire Input1 Mid1 Output1 @ 1 A treasure! Ques- Tionnaire Input2 Mid2 Output2 @ 2 Ques- Tionnaire InputX MidX OutputX @ x November 2007 5
A treasure! Ques- Tionnaire Input1 Mid1 Output1 @ 1 Ques- Tionnaire Input2 Mid2 Output2 @ 2 Ques- Tionnaire InputX MidX OutputX @ x November 2007 6
A treasure! All domains Same method Good metadata Coherence Social Skat CPR Person id: Person nummer Arbejdsmarked Uddannelse Spørgeskema Interview Combining across Sundhed Bolig & jord id: Eksakt adresse Arbejdssted id: CVR-nr. etc http://stats.oecd.org/wbos Matrikel BBR CVR November 2007 7
Users want to get quality data oecd.org Social Sundhed Bolig & jord id: Eksakt adresse Matrikel Skat CPR Person id: Person nummer BBR CVR Arbejdssted id: CVR-nr. Arbejdsmarked Interview Uddannelse Spørgeskema etc Billions of figures Fully updated Flexible query by theme by full-text search by stored query links to queries Other access methods from FAME & SAS SDMX web service November 2007 8
SIS Architecture 90 % of all OECD statistics The rest to come in 2008 Production Storage Dissemination Data Production (incl. StatWorks) OECD.Stat Corporate Data Warehouse Online access Books MetaStore Statistical Metadata XML XML CDs etc. November, 2007 9
The data model Multidimensional tables (cubes) each owned by one database administrator with metadata attached everywhere common dimensions country frequency time age sex... Arranged in themes and subthemes November 2007 10
SIS benefits Quality Harmonisation of concepts Improved statistical metadata Coherence of data and metadata across datasets Dynamic updating SDMX (ISO 17369) User friendliness One-stop database Combine data across themes Alternative outputs for different audiences Internal Efficiency November, 2007 11
November, 2007 12 Accessing OECD statistics
Access to OECD.Stat Access to the data warehouse restricted OECD officials government officials through OLIS Limited datasets on the Internet 3.1M downloads in 2007 of which 2.8M Internet use 70,000 from OLIS 140,000 internal OECD November, 2007 13
Accessing OECD statistics OLIS (with a logon) horizontal access to all databases and vertical access within each database targeted at users who have skill and time books and publications Free on Internet: Core Data key ready-made tables and sub-sets 500-1000 tables and sub-sets targeted at expert users and informed lay users Statistics Portal StatLinks e.g. Factbook 2007 p. 133, graph November, 2007 14
November, 2007 15 Most accessed datasets 2007 DATASET 2007 Country statistical profiles 2007 313124 Country statistical profile 2006 256269 1--Gross domestic product 170420 Composite Leading Indicators (MEI) 110138 Financial indicators MEI 96328 MEI Labour Statistics 86071 Main Economic Indicators: Archive data and revisions 66459 Reference Series 52844 Quarterly National Accounts 50911 LFS by sex and age 45426
November, 2007 16 Metadata: The key to understanding
Metadata principles 42 subcategories or metadata items how the metadata should be presented reuse of metadata : promote standardisation and coherence across domains supported by MetaStore November 2007 17
Access to metadata together with data in OECD.Stat well structured attachment levels independently and freely on www MetaStore Google for numbers easy to find deep descriptions When you found what you like >> Numbers November, 2007 18
November, 2007 19 Future
Develop infrastructures & products Develop integrated dissemination OECD Statistics Branding & standardising Core Data Develop Figures & Facts Diversify according to user groups Cross-domain products Dynamic graphics November, 2007 20
Tour Europe OECD The End November, 2007