Mining the co-existence of POIs in OpenStreetMap for faulty entry detection

Abstract

In recent years, more and more volunteers join crowdsourcing activities for collecting geodata which in turn might result in higher rates of man-made mistakes in open geo-spatial databases such as OpenStreetMap (OSM). While there are some methods for monitoring the accuracy and consistency of the created data, there is still a lack of advanced systems to automatically discover misplaced objects on the map. One feature type which is contributed daily to OSM is Point of Interest. In order to understand how likely it is that a newly added POI represents a genuine real-world feature, some means to calculate a probability of a POI existing at that specific position is needed. This paper reports on work in progress on a platform for analysing POI objects in the OSM database in order to find patterns of co-existence among features in close distance to each other. These patterns will improve current tracking and verifying systems and, thus, enhance positional accuracy of registered POIs in OSM.