Jrg Hackemann - Fotolia

Unicef uses data science to track refugees

United Nations agency for humanitarian aid for children and mothers in developing countries is turning to data science, marshalling expertise from companies and other corporates

Unicef, the UN relief agency for children and mothers in war zones, famine areas and other disaster regions, is starting to use data science to solve problems such as tracking refugees.

According to Natalia Adler, data, research and policy planning specialist at Unicef, this is just one example of how the organisation is looking to marshall expertise outside its own ranks to carry out its mission.

Adler said that some two years ago, the agency started to look at how the “data revolution could improve the lives of children”.

On the data science scene in New York, she came across Shawn Edwards, CTO of Bloomberg, whose firm was promoting a “data for good” scheme. In late 2015, Bloomberg agreed to fund a data scientist in residence at Unicef’s office in the city, and to open up some of its datasets for free. In this way, they mapped companies with child labour policies, and carried out network analysis to see how that carried through in those companies’ supply chains.

But Adler said the drive of companies in recent years to monetise their data had made it more difficult for organisations such as Unicef to get access to the data they would need to make dramatic progress.

Its strategy has been to launch what are called “collaboratives” with startups and other organisations, such as GovLab, a think-tank based at New York University. With GovLab, Unicef has developed a methodology that connects real-life socio-economic problems with data science and expertise that might help, especially from the private sector and academia.

Adler also said she was looking at the data science community in Scotland as another “ecosystem” to tap into. “It is small but growing and they know each other,” she said. “They have a willingness to venture into the social good aspect of data. I need brains, as opposed to just money.”

A Scottish government-funded innovation centre, the Data Lab, hosted a recent trade trip to New York with Scottish Development International, and one of the startups on that, Brainnwave, has become involved with the Unicef data science effort.

Unicef has also adopted, on Adler’s account, a business case-led approach, identifying problems to be tackled, then looking for the data that would help solve them. In recent months, it has identified nine problems to be solved, she said. 

Two of these are in Brazil – one concerning the usually high level of child killing in the country, and the other the unusually high growth of caesarean sections. The Unicef data scientists are using network and sentiment analysis to understand these phenomena better.

Read more about IT in aid programmes for displaced people

Unicef is working with Scottish data startup Brainnwave in a collaborative for one of its projects in Somalia, locating and tracking population movement in the country to enable Unicef to allocate resources and efforts to the areas in greatest need.

The UK Disasters Emergency Committee is currently putting a spotlight on Somalia and neighbouring countries, and has estimated that 16 million people urgently need food, water and medical treatment.

Some 60% of internally displaced people in Somalia are children, said Adler. When the Kenyan government threatened to close the Dadaab refugee camp last year – the biggest in the world, containing some 350,000 people – the need arose to understand where those people would move, to predict where goods and services should be sent.

Don Baker, founder of Brainnwave, brokered access to satellite data from, among other sources, Airbus to aid this project. Data scientists from the Turin-based ISI Foundation carried out research on the data, and continue to do so, pro bono, since the decision to close down the camp is still in dispute between the government and the Kenyan supreme court.

Another focus of concern is a zone on the Syrian and Jordanian border, where about 70,000 refugees are caught in a no-man’s land. Again, the idea is to use data science on satellite and other data to understand population movement in an area that it is very hard to monitor on the ground.

“The data collaborative is a new way of working for us,” said Adler. “We don’t come with all the answers, and we will never be in a position where we can hire lots of data scientists. It’s how Wikipedia works.

“There is no single answer to these complex problems, so we have to tackle them in a different way. It is collective problem-solving and, here, using data science to decode the complexity. Here it is data science, but it could be something else. Nor do you just solve a problem and say goodbye. It’s an ongoing thing.”

Adler will speak at Datafest 2017 in Edinburgh later this month. ..................................................................................................... ............................................................................................................

Read more on Business intelligence software