A MAP OF THE WORLD … DRAWN BY WORLD COLLABORATION. We all know someone – be it a friend, an acquaintance or even ourselves that have seen the famous Google Maps Cars as they’ve driven on a neighborhood street or a random location in the middle of nowhere. As we’ve seen it drive past, (some of us) have thrown our hands in the air or done something embarrassing in the search for StreetView infamy. What if we told you that those Google Maps Cars will no longer be needed and that you, as a regular person with a mobile phone, can help map the world and upload pictures to a community … and even get paid to do so? Since Swedish Mapillary was founded in 2013 by Jan Erik Solem, the company has done precisely that and mapped nearly 300,000 kilometers using images taken with mobile cameras and uploaded by members of the company’s community. The map service from Malmö is an example of what can happen when an entrepreneur dares to aim high .. and gives power to the people. Never underestimate a worldwide collaborative effort, just look at Wikipedia.
The prediction that Mapillary would become one of the few Swedish companies that would be able to compete with Google came in 2016 when users had uploaded 50 million images in the company’s database together with investments from Atomico and Sequoia in the amount of SEK 68 million. The company has come a long way since it was founded in 2013 by Jan Erik Solem who believed in making “a street-level imagery platform powered by computer vision”. Comparing the number of images from 2015 which was at 10 million, it was clear that the model was working and was expanding into new territory, away from the usual uploaders in Germany and the United States. The CEO mentioned that their platform was being used in a myriad of ways – “We have people who photograph hiking trails, bike paths and a guy from Holland who maps waterways from their kayak.” says Jan Erik Solem. [Source: Breakit]
But why use Mapillary when Google Maps and other major services exist? The power is in crowdsourcing. . .
The biggest and most crucial difference is the power of crowdsourcing. With thousands, eventually millions, of hobby photographers as users, Mapillary can map the world much faster and more extensively than an individual company’s employees (e.g. Google) can get close to. Some of those who take pictures are private individuals, other pictures are taken by organizations and companies that then use Mapillary to put the pictures together into maps. Another project that uses the Mapillary program is the Red Cross’s “Missing Maps”, where volunteers go out with go-pro cameras to photograph parts of the world that there are no maps of. In 2016, the company had mapped 1.2 million kilometers of land using user images – this is also where interest in the company rose significantly. The company took in SEK 68 million in venture capital from Niklas Zennström’s venture capital company Atomico and the US venture capital company Sequoia, which was used in further expansion. However, a break-through idea that will inspire further attention came this year and has placed the company in the tech-spotlight. Mapillary launched a marketplace for map data which will allow users to generate money. . .
This data marketplace is said to be one of the first in map data and will be used by other major players in their search to acquire data as necessary. We’re not talking about some small fish, but the likes of Uber and Lyft. In practice, the service functions as such that map companies and other actors can post different mapping projects in the platform officially called “Mapillary Marketplace”. After that, anyone can register as a Mapillary user, and get paid hourly to help complete the project by collecting the images needed with a specific app. All photos collected through the app are automatically uploaded to Mapillary’s database. There, the images are indexed, sorted and linked together using image analysis and deep learning algorithms. Accordingly, during processing, sensitive information such as faces and registration plates is automatically erased. Each pixel in the street pictures is given a label that tells what the photo contains. The Malmö company’s ai system has been trained to automatically identify 43 different types of objects in pictures – everything from traffic signs, to mailboxes and benches. [Source: NyTeknik]
Never underestimate the power of the crowd, especially when that crowd is getting paid. The dominance of crowdsourcing in recent times has proven the inevitable – the importance of simply providing a platform for users to contribute information goes a long way. Whether Swedish Mapillary will take the throne in this aspect is difficult to say, especially battling giants like Google on its journey. What we can say is that the new Mapillary Marketplace may drastically shake things up. From one of the oldest maps, the Imago Mundi from Babylonian times to the present – we have come a long way in projecting mapped information. Once mapping and mastering the dynamic world is complete – let us look towards other planets and celestial objects.
Evolvera – evolve in a new era