[This article was published in Position magazine’s April/May Issue – Out last week in print only. To read it in full please consider subscribing.]
Responsible for mapping earthquake damage after the 2011 Christchurch earthquake, searching for missing flight MH370, and most recently, assisting recovery in Fiji after tropical cyclone Winston, Tomnod has proven what’s possible when you bring together ordinary people and maps. Ahead of his appearance at Locate in Melbourne on 12-14 April, co-founder of crowdsourced mapping platform Tomnod, Luke Barrington, discussed with Position how we can expect Tomnod to evolve from here.
Your impressive TEDx talk in 2012 brought crowdsourced mapping to a wider crowd. Can we expect you to delve deeper for the geospatial crowd at Locatel6?
Absolutely. My work with Tomnod was really focused on making geospatial analysis tasks – like finding features in satellite imagery – as easy as possible so that it could appeal to the masses. More recently, we’ve been delving deeper into more sophisticated ways of extracting location intelligence using tools like computer vision, machine learning, spectral analysis, data visualisation and cloud-hosted workflows for crunching trillions of pixels. I’ll be talking about all kinds of “geospatial big data” at Locate16.
Do you think the crowdsourcing movement will continue to strengthen as computer literacy and accessibility expand?
In the 10 years that I’ve been involved in human computation, citizen science and crowdsourcing, I’m continuously amazed by all the new applications we’re finding for tapping the wisdom of the crowd. We’ve evolved beyond just the directed efforts like Tomnod (understanding satellite images), EyeWire (mapping neurons in the brain) or Galaxy Zoo (classifying astronomical data) to new, passive forms of crowdsourcing which filter the “data exhaust” the human race creates online. Now data scientists are mapping tweets, classifying photos and analysing online videos to extract new insights about how we live, work and play, both online and in the real world —and we don’t even have to do any work to create this data.
In which situations do you find the most interest from the public? Obviously the ongoing search for MH370 has garnered a lot of attention, but for what other areas might you expect to receive tremendous support?
Location is global but people still feel most affinity for events close to home. We regularly see campaigns on Tomnod get a huge response from local audiences, whether it’s Australians responding to wildfires in Adelaide, Argentinians searching for missing sailors off the coast of Buenos Aires or crowds in the US mapping changes in urban construction. Of course, our responses to major natural disasters like the Nepal earthquake always receive a ton of interest from the crowd, no matter where they call home.
With Tomnod having so much influence on real world situations, how can you decide what best warrants its use? How can you decide to focus on identifying swimming pools in Adelaide, rather than illegal logging in South America?
A big driver for us – and our crowd – is to work on campaigns where the outcome of our online crowdsourcing will actually have a real-world impact. Even though we can look anywhere, we can’t be everywhere so we rely on partners and customers to use the data we create to take action in the world.
(To read the rest of Luke Barrington’s Q&A, keep a look out for the April/May Issue of Position magazine. Out now in print.)