Pressing the crowd for data

September 21, 2018

With Facebook on the verge of making an initial public offering worth tens of billions of dollars, the value and power of social networking has never been more apparent.

But how can we capitalise on today’s wireless world to do more than just create online static? How can we harness the strength of its vast numbers – Facebook alone has more than 800 million active users each month – to do useful work?

These are questions that are likely to continue to produce new answers as technology develops.

For now, we in the computing and information sciences are curious about using crowd-sourcing – the practice of enlisting many people to engage on a specific task or problem – to tackle challenges and gather information.

To test some of the theories on social-network mobilisation, a team that includes several scientists at the Masdar Institute is lending its expertise to the MyHeartMap Challenge launched by the University of Pennsylvania.

The competition is simple: to track down as many automated external defibrillators (AEDs), the electronic devices that can treat heart problems, in the city of Philadelphia as we can.

The team that locates the most AEDs between January 31 and March 13 wins.

On our team are some winners of the pioneering Red Balloon Challenge, a previous crowd-sourcing contest in which researchers from the Massachusetts Institute of Technology used social networks to locate 10 weather balloons scattered around the US.

By harnessing the power of the masses, the challenge was cracked in just a few hours.

For this challenge, our team is spread around the world, with members in Abu Dhabi, San Diego, and Southampton, England – but none in Philadelphia. We will be testing the bounds of remote crowd mobilisation and data verification.

We will use social networks to pass the message on, to encourage people to report the location of AEDs and to verify other reports. If we win, the US$10,000 grand prize will be split among the people who helped. Give us a correct location, and you get a share. Pass the message on to someone else who gives us a valid tip, and you get a share.

We hope that incentive will help us to recruit as many people as possible to our cause and test some of our theoretical research on social-network mobilisation and incentivisation, as well as verification.

Verification is particularly tricky. One thing we learnt from the Red Balloon Challenge was that people lie or get things wrong. So the question is whether giving incentives for correct information can reduce this.

Finding ways to verify crowd-sourced information is crucial. If you are asking people to report the worst-affected areas in a forest fire, for example, you need to be sure the reports are accurate. And while contests such as MyHeartMap and Red Balloon challenges may seem purely academic and even frivolous, they can actually help save lives.

Having a map like the one that will result from the MyHeartMap Challenge, of most if not all of the AEDs in Philadelphia County, will be invaluable when people suffer heart attacks.

It, and projects like it, could easily save the lives of some of the 300,000 people who die in the US every year from sudden, out-of-hospital cardiac arrests. And that is hardly frivolous.

Dr. Iyad Rahwan is an associate professor of computing and information science at the Masdar Institute of Science and Technology.