Dr Andrea Aparicio Castro, University of Oxford
Department of Sociology (42-43 Park End Street) or MS Teams
Please join either in person or online. For in-person attendees, the talk will be preceded by a light lunch at 12.15pm.
Please email comms@sociology.ox.ac.uk with any questions.
Abstract
This talk presents a Bayesian framework for estimating international migration flows using imperfect and partially observed data. Migration statistics are often fragmented across sources and time, with substantial undercounting and measurement error, limiting their usefulness for understanding migration systems. Focusing on migration corridors between the EU-27 and the UK from 2011 to 2022, I show how traditional demographic data can be integrated with digital trace data to improve the estimation of migrant stocks and, subsequently, migration flows.
The approach proceeds in two stages. First, Bayesian hierarchical models are used to estimate observed and unobserved migrant stocks by combining census-based sources with social media data. Second, these stock estimates are used as inputs to derive consistent estimates of migration flows across corridors and over time. The talk introduces key concepts such as migration systems, corridors, and the stock–flow relationship, and illustrates how Bayesian modelling enables uncertainty quantification and comparability across data sources. I conclude by discussing implications for migration research and official statistics.
Biography
Andrea is a Senior Research Fellow in Bayesian Digital Demography at the Leverhulme Centre for Demographic Science (LCDS), where she joined in 2024. At LCDS, Andrea leads the development of statistical and computational Bayesian methods to estimate and nowcast populations in places affected by war and crisis, supporting international humanitarian response.
Andrea has a PhD and an Honorary Research Fellowship in Social Statistics from the University of Manchester. She is also accredited as a Teaching Fellow of the Advance Higher Education Institution. Her research focuses on developing statistical methods that integrate and combine traditional and new forms of data from multiple sources to estimate and forecast demographic events, with an emphasis on migration and mobility. She works primarily within the Bayesian framework, which enables her to correct measurement errors and systematic biases, as well as to impute and smooth out imperfect data.
