Earlier this year the Behavioural Insights Team, working together with Oxford’s Blavatnik School of Government and Harvard’s Behavioral Insights Group, hosted a one-day conference exploring the topic of data science and government. The conference attracted 200 participants from across government, industry, and academia.
In his opening remarks, David Halpern, BIT’s CEO and the National Advisor on What Works, observed that the world of data for government has transformed dramatically over quite a short period. Whilst David’s PhD thesis required weeks of work combing through hundreds of consultation documents in a single GP practice to collect data, our paper published earlier this year in the Lancet allowed us to analyse all of the prescriptions data from every GP practice in the UK over a two year period in a matter of hours. This was thanks to the Health and Social Care Information Centre (HSCIC) which was founded in 2013.
In the morning sessions, we heard specific examples of how data science is being used in academia, government, and industry to solve problems of interest to policymakers. Suzy Moat and Tobias Preis, directors of the Data Science Lab at Warwick Business School, presented some of their research on using various forms of online data to measure and predict behaviour. They showed how it is possible to use photos uploaded on Flickr to track travel patterns around the world and how an online game can help us better understand the relationship between living in a scenic environment and health. PhD students from the Data Science lab at Warwick also presented at the junior scholars session, showing how mobile phone usage data can be used to predict how many people are in a location much more easily than manual counting, with implications for resourcing of government programmes.
Aimee Murphy, Ben Howick and Jovita Lebednykaite from the Department of Education (DfE) shared how the DfE has been using predictive analytics to forecast events and better target resources. Frank Bowley and Adrian Jones from the Department of Business, Innovation and Skills (BIS) shared how they are are using data held by BIS and HMRC to get a better understanding of how young people move through education and into work.
Luca Aiello from Bell Labs outlined the Good City Life project which focuses on building sensorial maps – incorporating sounds, smells, and sights – of twelve cities. The data is generated from social media and aims to give more nuanced perceptions of urban environments. In the same session, Jeremy Reffin from the Centre for the Analysis of Social Media presented their in-house natural language processing tool – Method 52 – which can be used by non-experts to collect and analyse data from social media. Pete Burnap and Matt Williams, who direct the Social Data Science Lab at Cardiff University, also presented their tool, COSMOS which they have developed to make it easier for policymakers to make use of social media data in their work.
The afternoon sessions moved onto some bigger-picture discussions – with a series of short keynote talks and a panel discussion. Sir David Spiegelhalter, Winton Professor for the Public Understanding of Risk at Cambridge University, discussed some of the challenges of presenting data, such as health risks, to the public in a clear and unbiased way. Kate Glazebrook from the Behavioural Insights Team discussed some of the challenges she has faced in designing Applied, a tool to reduce bias in recruitment. For example – what can we do when the evidence strongly suggests that algorithms make better decisions than humans do, but people still resist using them? Perhaps the most important issue raised in the final panel discussion was by Professor Helen Margetts, director of the Oxford Internet Institute who posed the question: “what complex new ethical issues arise for a government with access to ever increasing amounts of data?”
At the end of the day, it felt like we were left with more questions than answers – perhaps a good starting point for next year’s conference!