On Tuesday 23 March 2021, the RSS North Eastern local group hosted a meeting on Bayesian Inference for Discrete-Time Spatial Epidemic Processes. The speaker was Dr Chris Jewell from Lancaster University. The talk took place remotely via Zoom with 44 attendees.
Chris's excellent talk described the predictive modelling his group are performing on the Covid-19 pandemic in the UK, which is one of several forecasts used in the UK government's pandemic modelling group, SPI-M. The UK epidemic has been characterised by regional outbreaks, often preceded by local infection 'hotspots'. Even with ongoing vaccination campaign, detecting hotspots is critical for early intervention to prevent larger outbreaks.
Chris described how the group approaches this with a stochastic spatio-temporal model based on case testing and human mobility data; the challenges such as considerable censoring in the data; and the need for a high performance computing pipeline to fit the model using Markov chain Monte Carlo methods.
The talk was followed by a question and answer session.