Title: How to combine conventional and unconventional data sources in official statistics: Some opportunities, challenges and an example using the semantic web.
Abstract: Official statistics influences and covers a wide variety of areas and is in a way the senses of a complex society and its decision makers. It contains statistics about economic activities such as the national accounts system, agriculture, trade and price statistics, there are social statistics e.g. about health, education and there are statistics about other topics such as the labour market, housing, tourism, environment and much more. Data for these essential statistical outputs are collected, processed, analysed and disseminated in different production processes with purposely tailored methodologies. Often, but not always, the responsibility lies with a National Statistical Institute (NSI) like Statistics Norway. The conventional data sources are Censuses, probability based samples surveys and use of administrative registers. However, under the pressure to modernize and to extract more value and better reuse of existing data, NSIs around the world are investing in ways to use other data sources. Data from satellites, social media, transactions and sensor signals are all under consideration. My talk introduces the opportunities and challenges that face NSIs in this work and I will illustrate with how techniques and methodology from the semantic web may be used.