Geoffrey NichollsView profile
Associate Professor
Geoffrey Nicholls is an Associate Professor of Statistics at the University of Oxford and a Tutorial Fellow at St Peter's College. He joined Oxford in 2005 from the Mathematics Department at Auckland University in New Zealand and served as Head of the Statistics Department from 2012-2015. His research focuses on developing new methodology and algorithms for statistical inference, with particular expertise in Bayesian methods. Prof. Nicholls' research interests span Bayesian inference, approximation methods and calibration, misspecified models and Semi-Modular Inference (SMI), ranking and partial orders, and Monte Carlo methods. His work demonstrates a strong interdisciplinary approach, with ongoing collaborations in archaeology, business inventory management, environmetrics, finance, genetics, history and philology, satellite imaging data, geothermal reservoir imaging, ecology, social network analysis, and historical linguistics. He has developed new methodology in Statistical Machine Learning and Computational Statistics since 2015. His recent publications reveal a continuing emphasis on advancing Bayesian methodology, particularly in areas of partial orders, semi-modular inference, and diachronic sense change modeling. His work shows strong connections between theoretical statistical development and practical applications across diverse fields, with a notable focus on historical and linguistic applications alongside more traditional statistical domains. Prof. Nicholls has supervised numerous doctoral and master's students, with recent graduates working on topics including statistical inference for partial orders, Bayesian methods for misspecified models, and scalable statistical inference. His teaching includes Bayesian methods courses at both undergraduate and graduate levels, reflecting his deep expertise in this area. As an active researcher with publications extending to 2025, Prof. Nicholls continues to make significant contributions to statistical methodology while maintaining strong connections to real-world applications across multiple disciplines. His work bridges theoretical statistical development with practical implementation across diverse scientific domains.
