At a time when weather forecasts steer military operations, geological cores guide clean energy investments, and border data shapes ministerial announcements, it’s little wonder that government now talks about data as a national asset.
That was the focus of an expert panel session at Think Data for Government in London (pictured). It brought together leaders from the Met Office, British Geological Survey (BGS), the Home Office and CACI to explore how government, defence and national security organisations can make data “secure, connected and ready for real-world decisions”.
From forecasts to impacts
Opening the discussion, Simon McLellan, head of data engagement at the Met Office, described why weather information sits at the heart of operational decision-making.
“You always have the weather with you,” he said. “The weather impacts all of our lives and livelihoods… What decisions are made by government, by businesses, by industry, and those industries can affect national infrastructure, transport, energy supplies.”
The Met Office holds “huge amounts of data that has happened, but also that prediction of the future”. Treating that information as a national asset means not just forecasting weather, but advising on consequences: “What the weather will do and what the weather means for you.”
Increasingly, that means experts sitting alongside decision makers, from local fire incidents to COBR. “We have contingency advisors… if there’s a chemical fire somewhere,” McLellan explained. “We have people going into Cobra meetings so that we can explain to the Prime Minister… what this weather situation might mean for them.” Operational staff also work “embedded in Heathrow” and “with national highways”, while uniformed colleagues support RAF and military planning “in theatre”.
McLellan noted the importance of data-informed decision-making rather than data alone. “We don’t expect people to have a PhD in geology or meteorology to understand weather data,” he said. “Our job is to translate scientific data into meaningful insights.”
Beneath our feet: data in physical form
Emma Bee, head of the national geoscience data centre at BGS, highlighted how foundational geological information underpins everything from groundwater safety to battery manufacturing.
“Rather than the atmosphere, we’ve got the lithosphere,” she said. “We apply that geological intelligence… to enable people to understand geology and what it means to them.”
One key lesson: partnerships matter. “A lot of decisions are made from geological information,” said Bee, including preventing subsidence on “London clay”, securing clean water and sourcing critical minerals internationally.
And unlike most digital datasets, geological evidence can be literally rock-solid – including the 600 kilometres of core that housed at the National Geological Repository. Environmental data can exist only in “physical form” and “there is work to be done even to get digital analogues”.
Data at the border
For Richard Appiah, head of data strategy at the Home Office, the urgency of accurate data couldn’t be clearer.
“We work in such a rapid pace of the political context,” he said. “An immigration officer needs… the right data to make a decision.” Whether assessing risk at the border or briefing ministers “at 12 o’clock… saying, ‘I’ve got a topic of announcement’,” delivery must be fast and right.
If you liked this content…
Migration data is often the subject of intense public scrutiny: “small boats… asylum seeking hotels… even issuing passports”. The question is: how can officials trust the data they are given?
“There is a huge cultural shift in terms of actually trusting data that is… siloed, and how do we get that into sustainable platforms?” said Appiah. Data literacy and “the human element” remain crucial: a dataset alone “doesn’t have a human at the end of it”.
Appiah recently spoke with the Windrush Commissioner about the “impacts of the Windrush generation… worldwide”. Human consequences demand human interpretation.
In practice, the Home Office often must provide a “single version of the truth”, even when “the data is not perfect”. That means “some friction, but it’s very much focused on business need,” said Appiah.
Trust and explainability in the era of AI
For Lee Griffiths, head of technology at CACI, the phrase “national asset” goes beyond policy language. “As ChatGPT told me,” he joked, “an asset is a valuable resource that will provide future benefit.”
What astonishes him is that those benefits rely on “zeros and ones fizzing around datacentres” ultimately affecting people’s safety. That places demands on data quality, privacy, explainability and security.
Encryption helps keep data safe, he said, but secure access matters just as much: “It’s no good just encrypting your data and then… it just sits there. You need to be able to share it.” He cited emerging technologies such as homomorphic encryption, zero-trust architecture, and differential privacy.
On AI explainability, Griffiths argued for subtlety. Scientific discovery like DeepMind’s protein-folding breakthroughs can succeed without full transparency, because “it wasn’t like… if this thing went wrong… someone’s life” was at risk. But in critical public sector functions, “I think that’s when you absolutely do need the expert in there.”
Models also “change as new data comes along,” making MLOps essential to ensure systems keep behaving as expected. “Deeply understanding why” a model degrades is part of operational trust, he said.
Valuing public data: more than monetisation
Asked how government should value its data assets, McLellan urged a nuanced approach: “‘Open data’ is a very clunky way of describing ‘this is open, off you go’.” The Met Office competes commercially in a global market, bidding to provide data services while maintaining public outputs like the BBC forecast. “We definitely got that understanding of the value we place on some of our data assets,” he said.
Bee added: “Someone has to invest somewhere.” Licensing value-added geoscience products helps keep national datasets current.
For Griffiths, the principle is simple: sometimes data drives decisions; sometimes expertise must override. “Particularly when… it’s making decisions automatically,” he said, “that’s when no – actually continue the expert in there”.
Across all organisations, collaboration is becoming non-negotiable. Whether through joint hazard planning or shared observatories monitoring space weather, the panel agreed that data gains its value when it is connected and used to protect people.
