370 likes | 385 Views
This presentation provides an overview of Eurostat and the European Statistical System (ESS), highlighting the data revolution and innovation in official statistical production. It discusses key challenges and research areas, including the use of multiple data sources, data mashups, and the need for tailored services for customers. It also explores the role of the ESS in addressing methodological challenges and leveraging big data and administrative databases. The presentation concludes with the communication revolution and the availability of new digital tools for accessing and disseminating statistics.
E N D
Data revolution and innovation in official statistical production: key challenges and research areas
Content of the presentation • Overview of Eurostat and the ESS • produced statistics • Data revolution • innovation in official statistical production • key challenges and research areas • Communication revolution
Eurostat Joseph Bech building, 5, rue Alphonse Weicker, 2721, Luxembourg
We provide high quality statistics for Europe ...making a difference in the ocean of information
The European Statistical System (ESS) A partnership of Eurostat and the National Statistical Institutes and other national authorities of the EU, the EEA and the EFTA countries
How does this partnership work? • ESS: • Harmonisation of methodologies, concepts and classifications • National statistical offices: • Collection of data • Eurostat: • Consolidation of the data • Production of European aggregates
Some things we don't do... • Economic and political comments (Spokesperson’s Service) • Public Opinion - Eurobarometer (DG Comm) • Economic forecasts (DG Ecfin) • Business/consumer confidence (DG Ecfin) • EU budget figures (DG Budget, DG Regio) • Tax rates (DG Taxud)
Data Availability Limitation of traditional surveys: • Increasing non-response rates • Concerns about response burden • Lack of flexibility and associated costs
Innovations and changes of the statistical production cycle • Use of multiple data sources • Data mashups • A new data "factory" • Data analytics services for "prosumers"
Use of multiple data sources • Extending traditional data sources to administrative and big data
6.Support & Coaching 4.Sampling frames COMPRENSIVE USE OF DATA & BURDEN REDUCTION 5. Pilot projects • ADMINISTRATIVE SOURCES: • Population • Health • Taxes • Social security • Census • Education • Unemployment • … … ADMIN
example Mobile network data forpopulation statistics (Belgium) Census (2011) Mobile phones (2015)
Multidimensional Poverty Index (Lighter colour indicates higher poverty) Poverty map estimated based on mobile phone data Poverty map in finer granularity estimated based on mobile phone data Smith, Christopher, Afra Mashhadi, and Licia Capra. "Ubiquitous sensing for mapping poverty in developing countries." Paper submitted to the Orange D4D Challenge (2013).
Data mashups Assembling and reassembling data from multiple sources Aim is to improve analytics
Data analytics services for "prosumers" Statistical organisations need to extend their products according to their users' needs: • Tailor-made services for customers • Data-driven surveys • Economic modelling • Forecasts and projections Change without modernising the production processes would not be effective
ESS domains DIGICOM BIGD SERV Europe 2020 Economic Governance & Globalization ESBRs ESDEN VALIDATION ADMIN ADMIN Economic and Social Performance Environmental Sustainability Business People's Europe Geospatial, Environmental, Agricultural & Other SIMSTAT/ REDESIGN Cross cutting
Challenge: Data integration • Integrating data coming from multiple sources collected in various ways • Will rely on DIGIT platformsand tools developed for exploiting new data sources • Fostering the development of methodological approaches to compile/describe sets of cross-domain indicators • representing and explaining interlinks
Challenge: Big Data and Administrative databases • Cleaning the data • Constructing the population weights • Sample recalibration techniques vs missing structural variables • Representing and explaining interlinks
Problems • BDATA: Self-selectivity bias of the sample • ADMIN datasets often suffer from lots of missing records and duplicates. • BDATA: suffers from high level Noise. • BDATA: often miss structural information, making linkage between datasets impossible. • BDATA: Spurious regressions
Opportunities • New Metrics • New Methodologies and algorithms • Integration matching techniques • New data collection - webcams
Communication Availability Limitation of traditional tools: • No timeliness • Limited value added to statistics • Only for specialised users • Lack of flexibility
Eurostat wiki Search all SE publications New section with easy explanations
Interactive tools Economic Trends Country profiles Regional Statistics Illustrated
Mobile apps Country profiles My Region EU Economy Eurostat quiz Available for iPhone, iPad and Android