We are pleased to be collaborating with Guest Editors at NOAA, the Met Office, the German Aerospace Center, the Climate Research Centre in Singapore and Oxford University on a Call for Papers on the topic of Environmental Informatics (deadline for submission 12 November) for Environmental Data Science, a new open access journal published by Cambridge University Press. In this blog, Samantha Adams and colleagues introduce a community initiative from the Met Office to promote data science.

Over the last 18 months the Met Office has put concentrated effort into supporting and promoting Data Science through various initiatives: a Data Science Strategic Action; incorporating a ‘Fusing Simulation with Data Sciences’ theme within its Research and Innovation Strategy and has recently formed the Joint Centre for Excellence in Environmental Intelligence (JCEEI) in partnership with the University of Exeter.

A key outcome of the Met Office Data Science Strategic Action has been the establishment of our internal Data Science Community of Practice (CoP) – a (currently virtual) community for sharing knowledge about Data Sciences. The CoP launched in April 2020, running a series of tutorial sessions aiming to establish a base skill level in the community at which future sessions can be pitched. After these initial sessions, we have had regular meetings with talks from internal and external speakers about theory, applications and projects involving Data Science and Machine Learning. We have recently also started to run more hands-on learning activities like internal Kaggle-style competitions. These serve to explore Data Science problems, drawing in a wider range of skills from across the organisation to tackle the problem being considered, but also for a development opportunity for those participating. The curated dataset and code that is used then form a tutorial that can be used for training after the competition has run. We are also establishing additional activities, like a journal club to keep on top of the latest research and a mentoring scheme. Currently there are over 140 members of our Data Science CoP!

The Met Office is not the only weather and climate centre that has, or intends to, establish a Data Sciences Community to promote knowledge sharing within their organisation and we think that collectively we could do much more than any individual organisation. Therefore, the Met Office has proposed and launched a ‘Network of Data Science Communities’ in order to share relevant training material, experiences and best practices around applying Data Science and Machine learning to Earth Sciences and beyond. As a starting point, we have put together two mechanisms for sharing. The first is a GitHub repository that we will, at least initially, primarily use the wiki page as a portal for links to resources created by member organisations. Secondly, we have created a SharePoint site and Office 365 group to facilitate external people to join our Data Science CoP sessions either as attendees or speakers. The SharePoint site will also enable sharing of documents and training material.

The Network of Data Science CoPs is still in its early stages, and we welcome new members, contributions of useful material and ideas for how to collaborate more effectively!

If you are interested in joining our community, please email stephen.haddad@metoffice.gov.uk and samantha.adams@metoffice.gov.uk for more information.