“It’s not about what data can do for diplomacy. It is how diplomacy can possibly remain relevant unless we embrace data.”

So says Graham Nelson, the founder of the UK Foreign Office’s Open Source Unit (OSU). He is fresh from delivering a seminar on data-driven policymaking at Singapore’s Ministry of Foreign Affairs.

It’s been a long day, but he becomes visibly more animated when talking about his work: helping governments around the world use data to solve their most defining challenges. “I am really excited by the potential for data to do so much good,” says the mathematician-turned-diplomat. He shares how data is an indispensable tool for governments today, and how it can help agencies examine the impact they are really making.

Crisis response

The OSU was set up in 2016 to give the UK government greater clarity, confidence and competitiveness in an increasingly complex and uncertain world, according to Nelson. At the same time, there is the foreign aid element: “Data can be a huge enabler in terms of commitment to aid, and how the UK government supports equitable development.”

The unit has developed a tool that automatically ingests and translates hundreds of thousands of multilingual news sources into English, creating a real-time feed of updates on a dashboard. “We are essentially getting a flash report from the region, including very local publications, not just your big and well-equipped news outfits,” he shares.

“Previously, diplomats had to spend lots of time reading local news, speaking with local partners and engaging face-to-face,” Nelson continues. “To be able to combine that now with the insights we gain from conversations happening online, we have a much better sense of the world.”

His work also revolves around using data to solve complex problems that governments face, like in disaster recovery. It is all about shedding light where there was previously no visibility, and give “a much better sense of what is happening” – especially during a crisis.

Deployment teams are able to rapidly draw out themes across countless articles, reports and online conversations that would be humanly impossible to read in a short amount of time. “Even when we don’t have individuals in situ, we can help to understand where we should be deploying our aid and crisis response teams,” Nelson explains.

This proved crucial during the recovery of the 2017 hurricane in the Caribbean, according to Nelson. Communications were down, infrastructure was poor, and challenging weather conditions hampered aid efforts. The tool helped to provide much better awareness of the situation.

Data for global governments

It certainly helps that governments today “have got much better access to commercial satellite data and meteorological data than we would have had before”. “There are some really easy ways that countries thinking about setting up on this journey of using data can start,” Nelson points out.


“In many ways it is going to be the Asian century, so I think if anywhere is going to be impacted by data, it’s going to be here first.”
Data can make an impact in “really tangible ways” in Asia Pacific, in particular, he believes. “In many ways it is going to be the Asian century, so I think if anywhere is going to be impacted by data, it’s going to be here first,” he remarks.

International collaboration is key to what the unit is trying to achieve, and precisely why Nelson was visiting Singapore in the first place. “The challenges that we face today are too big for us to be basing our response to them on assumption, or historical traditions,” he emphasises. “They’re also too big for us to tackle individually.”

Tracking extremism

At the same time, data needs to be put into a human context for it to be useful, says Nelson. The unit has behavioural scientists looking at the data to make sense of the trends and analyse the underlying behaviours.

The unit tracks “instances of extremism” to understand global trends and decipher what that could mean for populations vulnerable to it. This way, the UK government can understand how to use campaigning and diplomatic engagements to “try to influence, and to deradicalise, or to reach out with different messaging” – but within clear legal and ethical frameworks, he emphasises.

The UK government’s definition of extremism includes extremist ideology that can lead people to violence, which divides communities, opposes “fundamental British values” and which says that “‘the West’ is at war with Islam and that it is not possible to be a true Muslim and to live an integrated life in the UK”.

Analyses of global trends can also tease out whether internet searches of topics around extremism mean that the searcher is sympathetic, researching the topic, or simply curious. “That body of work that we do is a genuine collaboration where behavioural sciences is absolutely as valuable, if not more so, than the underpinning data,” Nelson notes.

Three tips on using data

Nelson is firm in the belief that good development policies require good data. He recommends that governments first identify the problem they are trying to solve, and along the way, bust any preconceptions of what they think the issue is.

Next, it is a question of whether quality data to help them solve the problem currently exists. “And if not, how could we look to generate it?”

And while it may be a bitter pill to swallow, data can help governments to examine their work very closely, and see if they are actually making an impact or not, Nelson continues. It could be a completely different picture than first thought, revealing that a lot of the work that any organisation does “is probably of limited impact”, he says.

These insights help governments and agencies cut out wasteful activities and become more impactful. The UK Foreign Office is now “much more rigorous in monitoring and evaluating our work”, Nelson explains.

The sun is now low on the horizon, and warm light filters through the blinds of the conference room. As we wrap up our chat, Nelson notes how he wants to share the OSU’s experiences in the hopes that other governments around the world also see the incredible value of data.

“Essentially it’s not a question of whether to do it,” Nelson concludes. “It’s a question of whether we can afford not to.”