Deep Data and Shared Computation: shaping the future Trusted Smart Statistics

The role of modern official statistics is to provide a quantitative representation about the state of the society and of the economy to policymakers, citizens and economic actors. Official Statistics is intimately connected to democracy1: it delivers the same information (“statistics”) openly, to everybody at the same time. A system of legal, organisational and methodological frameworks is in place to ensure that the whole production process, from data collection through processing and dissemination of the final statistics, is independent from other public authorities and private entities [2]. On the operational level, the production of Official Statistics is based on the model exemplified as depicted in Figure 1(a). The Statistical Office (SO) collects the input data centrally, processes them internally and finally publishes the information produced in output, i.e., the statistics. The quality of the final output depends jointly on (i) the quality of the input data and (ii) the quality of the processing method (algorithm). The ultimate goal of SO is to deliver output information with guaranteed high quality, including relevance and timeliness, and in the traditional model SO have full control over the design and execution of the processing method. They also have direct control over the collection of input data in the case of survey and census data. As for administrative data, they are collected by different public authorities and for purposes other than Official Statistics, however SO still retain accurate knowledge about the collection process, sources of errors etc.