Boundary estimation and social frontiers: A story of interdisciplinary collaboration
Speakers:
Bio: "Gwilym Pryce is Professor of Urban Economics and Social Statistics at the University of Sheffield where he holds a joint appointment between the Sheffield Methods Institute and the Department of Urban Studies and Planning. Gwilym is Director of the ESRC/Nordforsk Life at the Frontier Project researching the social mobility of migrants and the impacts of residential segregation. Gwilym is also the Sheffield Director of the ESRC Centre for Doctoral Training in Data Analytics and Society, an interdisciplinary PhD programme. Gwilym's research interests are primarily in social segregation and inequality, including the ways in which neighbourhood deprivation and social fragmentation influence patterns of crime and anti-social behaviour."
Bio: "Duncan is a Professor of Statistics in the School of Mathematics and Statistics at the University of Glasgow, and his research interests focus on developing models for spatio-temporal areal unit data, particularly in environmental public health applications. In addition to numerous published research papers and grants in this field, Duncan is the author of the well used CARBayes and CARBayesST R packages for spatial and spatio-temporal areal unit modelling. Additionally, he hosted the 2019 GEOMED conference in Glasgow, which is the premier interdisciplinary meeting focusing on spatial health data modelling. Finally, Duncan has ongoing collaborations with Public Health Scotland and Public Health England where he has honorary appointments, and has previously taught numerous short courses on spatial data modelling to researchers and public health analysts alike."
Bio: "Kitty Meeks obtained her DPhil in Mathematics from the University of Oxford in 2013, before taking up a postdoctoral position at Queen Mary University of London. She moved to the University of Glasgow in 2014, first as a Lecturer in the School of Mathematics and Statistics, before moving to the School of Computing Science where she held a Royal Society of Edinburgh Personal Research Fellowship (""Exploiting Realistic Graph Structure"") from 2016 to 2021. She currently holds an EPRSR Fellowship entitled ""Beyond One Solution in Combinatorial Optimisation"" (2021-2026). Her research focuses on techniques for the design of algorithms that are provably both correct and efficient, and on the application of these algorithms to many other areas, ranging from Social Network Analysis to Precision Medicine."
‘Social frontiers’ are sharp changes across space in the religious; social or ethnic composition of residential areas that mark territorial boundaries between neighbouring communities. These frontiers potentially exacerbate crime and reduce wellbeing. The challenge for social scientists has been to estimate these boundaries in a way that allowed them to be “open” rather than “closed”—a neighbourhood can have a sharp social boundary along one section of its perimeter but gradual transitions elsewhere—and to account for random variation. We realised that Bayesian spatial statistical methods being developed at Glasgow University for disease mapping could provide a solution. This led to a co-authored journal article estimating the impact of social frontiers on crime. Since then the interdisciplinary collaboration has widened further. Working with a computer scientist we are currently drawing on cutting-edge developments in graph theory to optimise the accuracy of the social frontier estimates. It’s a story of how social science can advance through interdisciplinary collaboration.