JUNE

JUNE builds a virtual population of a country from census data at fine geographic detail, with each person given an age, sex, ethnicity and socio-economic indicators.² An infection then spreads through the groups they spend their days in: households, schools, workplaces and other social activities.² How people mix within each group follows social mixing matrices.² Diseases are described by a set of parameters, so the framework is not tied to one virus.² It is named after June Almeida, the Scottish virologist who first identified the coronavirus group.¹

The code is kept by the IDAS-Durham group on GitHub, and the framework was described in Royal Society Open Science in July 2021 by Joseph Aylett-Bullock and thirteen co-authors.^1,2 That paper applied it to COVID-19 in England and reports that its first outputs reproduced hospital admissions and deaths at national and regional level and by age group.²

It installs from PyPI as june, and a script downloads the input data for England.¹ That data mixes several Office for National Statistics datasets and may contain modifications; the original sources are cited in the release paper.¹ Running it needs an MPI installation, the software that splits a large run across many processors, and the README carries notes for two computing clusters.¹ A quickstart notebook introduces the model.¹

Development has slowed: the last tagged release was version 1.2.0 in October 2022.¹ Using it for a country other than England means building a new population from that country's census first. The code is GPL-3.0.¹