EIONET European Environment Information and Observation Network http://www.eionet.eu.int/ Version 0.9 2004-11-10 * * * Quality assurance of Eurowaternet data Presentation for the EIONET Water Workshop Budapest, 11-12 November 2004 Hermann Peifer, Project Manager EIONET Data Flow
2 Quality: Models, concepts and frameworks Concerning quality in a broader sense: ISO 9000: A family of International Standards for quality management http://www.iso.ch/ TQM: Total Quality Management search on google.com: 472.000 hits search on amazon.com: 5.485 books EFQM: Excellence Model of the European Foundation for Quality Management http://www.efqm.org The word quality might have been overused in the past years. A too generic definition bears the risk of loosing focus.
3 Data quality frameworks Concerning data quality in the in sensu strictu: ESS: Quality Declaration of the European Statistical System: http://amrads.jrc.it/WPs%20pages/Quality/Documents/LEGsummary.pdf IMF: Data Quality Reference Site of the International Monetary Fund (IMF) http://dsbb.imf.org/Applications/web/dqrs/dqrshome/ MIT: Total Data Quality Management Program (TDQM) http://web.mit.edu/tdqm/www/index.shtml/ FAOs approach to data quality evaluation and monitoring http://www.fao.org/faostat/quality_en.asp OECD initiative on environmental data quality http://www.oecd.org/
4 What is data quality? Data quality is a measure of the degree of usefulness of the data for a specific purpose. Data quality indicators are qualitative or quantitative descriptors of data quality.
5 Data quality management in EIONET: Too little, too late? Data quality management is happening in EIONET. Much of the QA/QC work is carried out by Topic Centres. The level of available QA/QC documentation is improving. There is a clearly a lack of visibility for this area of work. An overarching framework for EEA/EIONET is missing. Annual Management Plan 2005 1.4.2 Maintaining and quality assuring priority data flows To provide an overview of the quality of the data provided by the member countries by documenting QA/QC processes country by country, in order to improve the overall quality of data.
7 EEA/EIONET data and information flow Data collection Aggregation Assessment Presentation Dissemination Flow ReportnetEEA/EIONET Assessment processes Information processes Tools to serve EEA, MS, others EEA/Users
8 Reports EEA NRC NRC NFP ETC Data quality: Where it matters most Reportnet Data collection EEA/EIONET Data processing EEA/Users Data dissemination Q1Q2
9 Data quality dimensions Representativeness Accuracy Punctuality
10 Which dimension is most important? Relevance Timeliness Comparability Accuracy Accessibility Completeness Many say: Relevance Another popular choice: Timeliness The classical one: Accuracy A smart answer is: All My personal preference:...
11 Guiding thoughts on measuring data quality You can not measure what is not defined. You can not improve what can't be measured. Focus on understandable data quality dimensions. Easy to measure, preferably via simple automated analysis. Providing the right level of information to data providers (Q1) and end-users (Q2) The aim: simple, regular, quantitative and user-oriented data quality reporting
16 Scoring criteria for Eurowaternet data collection 2005 Response to data quality questionnaire
17 Data quality in EIONET: Summary Data quality in EIONET needs: –More documentation –Continuous improvement –More visibility Topic-specific QA/QC procedures have to be embedded into a data quality management framework. The overall aim is (preferably) quantitative data quality reporting which is regularly updated in order to show trends over time. EIONET data quality reporting (mainly targeted at data providers, i.e. Countries) has to be complemented by data quality information targeted at end-users of the data. EEA Data Quality Framework QA/QC Euro water net