In the video below you can see the winning hacker, Michael, using our Watson IoT platform and the Watson Visual Recognition service. Michael Hsu finds Pokemon with Watson’s help Using the Watson Visual Recognition API Michael created a hack that takes periodical screenshots, identifies the Pokémon characters in them, and alerts other users to where the characters are. “I saw the Watson Visual Recognition API and wondered, what if I could get Watson to play Pokémon Go for me?” This weekend saw Michael Hsu, a front and back-end developer and part-time university lecturer in California, win the Best Use of Watson challenge at the AT&T Shape Tech Expo Hackathon in San Francisco. That’s where IBM’s Watson, and Michael Hsu come in… As the game has taken over, so have the range of ways in which players can out think Pokémon. The Pokémon Go game, using augmented reality and GPS, allows players to capture, battle, and train virtual creatures, called Pokémon, who appear on device screens as though in the real world. Pokémon Go has quickly become one of the most used apps beating the previous record held by Candy Crush Saga. The example map even lets you fly a small bat around the world on the map.Watson API is a winner with a new Pokémon hack
This Mapbox blog post provides a little example of a Pokemon Go themed map created using Mapbox Studio. If anyone is interested in creating map-tiles in the same style as Pokemon Go they could do worse than to have a look at Mapbox's Design Your Own Pokemon Go Map. PokeMapper does seem to have at least made a half-hearted effort to copy the general colors of the PokemonGo map tiles - but really hasn't got the colors right. In fact it uses very nicely designed custom map markers, which feature images of the Pokemon characters.
Pokecrew seems to be the only third-party Pokemon map to have given much thought to the design of the map. In terms of aesthetics all the crowd-sourced maps (bar one) appear to be very basic in design - presumably in the general rush to get to the market. It is possible therefore that PokeVision might stop working - if Niantic get upset with PokeVision or make changes to their internal API. I presume this is an undocumented API (and therefore could be against the Pokemon Go TOS). PokeVision uses the Niantic API to find the locations of Pokemon near your location in real-time.
Each Pokemon shown on the map also includes a timer which displays how long the Pokemon will remain at that location. It actually shows you the location of all the Pokemon near you in real-time. PokeVision does not use crowd-sourcing to find nearby locations of Pokemon. If you don't want to rely on the power of the crowd then you should use the PokeVision map. The quality of the data on the crowd-sourced maps are always going to depend on the number of active users and people submitting locations to the maps. Most of these maps appear to use crowd-sourcing to collect the data from Niantic Labs and Nintendo's 'augmented reality' mobile game. In light of this success lots of interactive maps are popping up promising to show you the locations of Pokemon characters, gyms and PokeStops. It is undeniable that the biggest mapping story over the last couple of weeks has been the huge success of Pokemon Go.