I am a freelance chartered statistician based in Hampshire, UK. I provide training courses, coaching and consultancy. My expertise focuses on Bayesian statistical models, evidence synthesis, data visualisation, health service evaluation, and communication to policy makers.
My clients have included the Cabinet Office, Harvard Medical School, the World Bank, The Economist, and the Royal College of Paediatrics and Child Health. I am also a Senior Lecturer on Kingston University's MSc course in Data Science every Autumn.
You will find my contact details via the menu above. Please note I am fully booked through 2024 calendar year and have no availability for new freelance work.
My first book, "Data Visualization: charts, maps and interactive graphics" was published by CRC Press and the American Statistical Society in 2019. The next is "Bayesian Meta-Analysis: a practical introduction", co-written with Professor Gian Luca di Tanna, scheduled to be published by CRC Press in 2025. We have started an online forum and blog, "the Bayesian Meta-Analysis Network" at bayesian-ma.net.
I am also the organiser of the Winchester Data Analysis Meetup, where everyone working with data in the local area can meet, share ideas and build a community.
I programmed the StataStan interface to Stan software for Bayesian inference, and have an ongoing project to create functions/commands for kudzu distributions (see the link above) in R and Stata.
A list of publications — peer-reviewed papers, commissioned articles, posters, talks, reports — is available via the link above.
I started my company, BayesCamp, in 2017. Prior to that, I worked on:
I was a member of the Royal Statistical Society's statistical computing committee, and NHS England's National Advisory Group for Clinical Audit and Confidential Enquiries.
My first degree was in mathematical sciences (first class honours), followed by a diploma in statistics, and masters in medical statistics. I will be examined in June 2024 for a PhD by Publication ("Applications of Bayesian latent variable models to challenges in health and social care data").
Most training in data analysis is not as effective as it could be. I believe that we could do a lot better by helping learners to think carefully about who their audience is, and what the audience need from the analysis, as well as communication skills. Analyses should then be tailored to fit the requirements, rather than following statistical rituals.
I use an eclectic range of statistical methods: Bayesian, frequentist, likelihood and machine learning. I think that the right tool should be used for each job. My personal philosophical standpoint, which emphasises justified eclecticism, is here.
I keep active in research and scholarship concerning: meta-analysis, statistics teaching, performance indicators, philosophy of statistics, Bayesian models, and statistics communication and visualisation.
This summer, I will be at the following conferences: ISBA Venice, StataConf Portland, JSM, RSS Brighton, StataConf London. Come and say hello. I might be presenting, or sharing a poster, on kudzu and non-parametric high-dimensional Bayesian updating.