Abu Dhabi: Social media users are helping humanitarian organisations better respond to natural disasters by sifting through volumes of big data, was the theme of a lecture at New York University Abu Dhabi.

Assisting humanitarian organisations cope with an overflow of information thanks to social media, digital humanitarians are being increasingly called on to help sift through large amounts of data.

The concept is a form of crowdsourcing, except in this case the crowdsourcing is done over the internet with users going through tweets, pictures, and videos of a natural disaster. As users comb the massive amounts of data, they pinpoint the most relevant information that will be of use to organisations looking to form a response to a humanitarian crisis.

Patrick Meier, who wrote the book Digital Humanitarians, has helped develop MicroMappers, a website where anyone can join up to become a digital humanitarian.

“One example of the work we did was after the Nepal earthquake. We went to our email list of volunteers and began to mobilise them because we were overwhelmed with the information we had. One task was set for volunteers to go through all the tweets and identify urgent needs. Another task was for volunteers to go through pictures on social media and news media looking at the infrastructure damage and to map its location,” said Meier in his lecture.

“For pictures we use photo clicker on our website which is a very simple application. The user looks at the photo to assess the damage and clicks on either red, orange, or green, with each colour symbolising the level of urgency,” he added.

Meier went on to explain that four to five volunteers assess each image that is classified as urgent to reach a unanimous decision, after which they go on to map the location of the picture.

While crowdsourcing has been very beneficial, the next step is to combine it with artificial intelligence, according to Meier,

“The key to moving forward is getting help from artificial intelligence, so you can say it’s a hybrid solution, human-machine collaborations. Humans are good at certain tasks, while machines are good at certain other tasks, so what we need is to bring them together to solve the big data challenge,” he said.

One such system has already been developed, the Artificial Intelligence for Disaster Response (AIDR). As the user classifies tweets, the AIDR is then able to identify the relevant tweets during a natural disaster.

Unmanned aerial vehicles (UAV) are another major solution that is currently being used as well,

“Usually after a disaster a field-based survey is done by humanitarian organisations which is incredibly time-consuming. With aerial vehicles they are able to go up in the sky looking down and survey the damaged area, collecting data and taking pictures. The aerial survey is able to identify which houses have been fully destroyed, partially destroyed, or largely intact,” Meier said.

The images produced by the UAVs can also then be accessed by digital humanitarians online, going through the imagery to try and find areas that have suffered the most damage and in need of urgent help.