We're running challenges around the world to create fresh coverage in selected locations and explore the potential for targeted image collection. The goal is to complete the map.
In May this year we launched the concept of #CompletetheMap challenges. Since the early days of Mapillary, we've heard of interest in tasking systems, leaderboards, and other tools that connect local communities and encourage fresh coverage. Giving people better indicators of where coverage is needed and where others in the neighbourhood are mapping is an exciting prospect.
More recently, we started thinking about how this could work in practice and where we should test it out. Chris was able to put together a very functional map and leaderboard that track contributions over the course of the challenge. In search for a trial city, we looked at contributions across the world. Berlin quickly emerged as a favourite with a high number of contributors in a relatively small geographic area. There was also a need for fresh imagery throughout central Berlin.
With the basic idea and structure in place, we began experimenting with a grid, working out ways in which we can track progress throughout the challenge. The idea of zones seemed very compelling, the intention being to start with an incomplete map that contributors could cover, unlocking zones as they do so. The grid would start out red, representing incomplete coverage. This would transition to orange as progress improved with green the ultimate goal—a complete zone.
With that, we launched #CompletetheMap Berlin, not really knowing what to expect. Here's what happened.
As you can see, on day 1 we went from a blank slate to imagery throughout Mitte and Kreuzberg. The massive jump on the first day can be attributed to Moenk on his motorbike, and Holgerd walking around on foot south of Landwehr Canal. Over the following days of the challenge, others joined in and helped fill in zones throughout the city. By day 18 there was not a single uncovered zone in the grid. This is an impressive achievement by Berlin contributors and a positive indication towards the role that tasking and gamification can have on imagery collection.
Moenk's motorbike setup
To keep the momentum alive, we have launched challenges in Budapest and Seattle, with more localities on the way soon. Budapest has a lively Mapillary community and community groups such as Járókelő. Járókelő crowdsources information on potholes, lights that aren't working, zebra crossings that need repainting, and so on. It's a vital source of information for the government and helps them stay up to date on infrastructure issues they may otherwise miss. It's early days, but exploring how street-level imagery can be an integral part of this process is something we're really excited about.
In Seattle we are working with GIS Master students at the University of Washington who will be using Mapillary data as part of their final year capstone project. To coincide with the project and to provide them with fresh imagery of the Seattle area, we have launched a challenge centred around the vibrant and densely populated Capitol Hill area. It will run until July 30 so there is ample opportunity to take an area with relatively little coverage and turn it into one the best mapped localities globally.
Some challenge tweaks
One change we have made from Berlin is to reduce the number of zones, making them larger instead. This makes for a more competitive challenge, at the same time making it simpler for both participants and yours truly to track. We've also changed the way each stage is calculated. Red is now anything less than 50%. Orange remains until the zone is 100% covered, in which case it becomes green. This will make the progress from each stage more significant and, hopefully, encourages mapping of smaller tracks and paths, not just the main roads.
Think of this as one big experiment that we are iterating upon every challenge. Your feedback is most welcome via all the regular channels. If you'd like to take part, nominate your city or speak to us about forking the code base we used to create the challenges.