National systems of official statistics are expected to provide governments, businesses and the public with data about the economic, demographic, social and environmental situation. Digitisation of data collection for official statistics is mooted as having a major potential impact on society. Access to big data for official statistics is currently stuck in trying to make more use of administrative data, including joining up data from different sources. However, there are truly big data science initiatives, building on data from satellites, mobile phones and social media. The United Nations 2030 Agenda and its Sustainable Development Goals commit governments to the production of new official measures of societal wellbeing and progress while recognising that such measures will only be deliverable through a data revolution. Effective government use of new data sources has the possibility of creating the ultimate evidence base for policies intended to improve lives. But will that happen? The information space is increasingly crowded with other players, some of whom draw on open data. We will explore how big data is envisaged to work in official statistics and start to assess the challenges to it working as intended. We conclude that technical developments need to be accompanied throughout with greater attention to the marketing of official statistics and engagement with users.
Paul Allin is a visiting professor in statistics, in the department of mathematics at Imperial College London. His research interests are the measurement of national wellbeing and progress, and the use of these measures in politics, policy, business and everyday life. He also lectures on official statistics and chairs the UK Statistics User Forum and the Advisory Panel of the What Works Centre for Wellbeing. Previously, Paul spent forty years as a professional statistician, researcher and policy analyst in various UK government departments and agencies. He set up and directed the Measuring National Wellbeing programme at the UK Office for National Statistics.