190 likes | 200 Views
Explore the use of administrative data sources in the Netherlands for statistics, including primary vs secondary sources, open data and big data, and the classification of administrative data sources.
E N D
Admin data The use of administrative data sources in the Netherlands May 12, 2016 Otto Swertz
What are administrative data? • Primary source vssecondary source (collectedby the statisticalauthority or collectedelsewhere) • Data fromanadministration or from the administration? • Only data on more thanone company? • Microdata or aggregated data? • How do open data and big data fit in?
What are administrative data? Ourchoice of definition Adminstrative data are all data which are collectedandmanagedbyanotherorganisation, but are usefulforstatistics. We call thissecondary data sources todistinguishthemfrom the primary sources.
Used admin data in the Netherlands • Grid operators for gas andelectricity. • National geologicalinstituteforminingactivities. • Subsidy register forrenewables. • Trade statistics. • Market authorityforgridlosses. • Motor vehicle registration. In development/wishedfor: • Registry of solar panels • Gas usedfor CHP (no energy tax on it) • Access dutyfor transport fuels
What is in the statistics law • It says more or less: we can get all data which are somehow obliged by the government. If they are public, then the organisation will have to deliver. If they are private, they have to be added to the law. • Client files ar held privately and are taken up in an act. • For every new dataset demanded by the government, but held privately the minister will have to add such an ‘act’.
Best paractice: using client files • Dutch Energy Balance (NEH) • Manufacturing industry & energy companies: questionnaires • Agriculture/homes: externalsources • Energy consumptionfor services sector? Black hole! • Solution? Requestfrom the energy distribution companies theiradministrative data withclientconnections: the “client files”.
Why use client files? • Client files contain records withall gas andelectricityconnectionswithin the Netherlands • 11 regionaland 3 nationaldistributioncompanies • More than 8 millionelectricityconnections • More than 7 million gas connections • No administrativeburden on respondents • Usabilityforstatisticalpurposes?
What’s in the files? • Homes / businesses / pumping-engines / cell phone towers / overhead contact wires (train), etc. • How to identify connections? (EAN code) • Records contain address data: postal code, house number, extension • Linking to other administrative data with known units • Distinction between houses and businesses
Electricity consumption Information and communication sector
6000 5000 4000 (kWh) electricity 3000 2000 apartement Row house 1000 detached Corner house Semi detached Semi- detached Corner house 0 detached Row house apartement One-person households Single parent household Married couple without children Unmarried couple without children Other households Unmarried couple with children Married couple with children Client Files households Bouwperiode 1960-1980
Specific energy consumption Natural Gas Government services Commercial services
Concluding remarks • Client files: very good source for energy statistics! • Solve linking problems and connect with other files • Solve assignment problems (how to cope with differences among registers) • Use spatial information (GIS) • May provide solution for assignment (block heating, greenhouses) • But: quality of individual records • Plausibility check • Special attention for (too) big mutations: error?
It brings opportunities • Use for regional climate actions • For determining specific energy consumption • For classifying neighbourhoods with ‘succes with energy saving projects’