Geodata Classification for Automatic Content Creation in Location-Based Games.

The ubiquity of smartphones with integrated positioning systems made it possible to develop location-based games playable by most people without requiring additional equipment. Popular location-based games like Ingress or Pokemon Go have demonstrated the public interest in this genre and studies indicate that playing such games has a positive health influence related to the players' increased movement. A big development challenge for these games is the content creation. Manually selecting points of interest (PoIs) of the real world for a game is time-consuming, expensive and often results in heterogeneous distributed PoIs with a low concentration in rural areas. In this paper we present a system that uses georeferenced data from open available sources to generate a collection of PoIs usable for location-based games of free definable target groups with the goal of providing a comparable game experience everywhere. The content creation algorithm in our approach is fully parametrized allowing for individual configuration for desired PoI criteria. Our evaluation shows that our system can be used to automatically select well distributed relevant PoIs for densely populated areas as well as for rural areas.