Data Quality Assurance for Volunteered Geographic Information

The availability of technology and tools enables the public to participate in the collection, contribution, editing, and usage of geographic information, a domain previously reserved for mapping agencies or companies. The data of Volunteered Geographic Information (VGI) systems, such as OpenStreetMap (OSM), is based on the availability of technology and participation of individuals. However, this combination also implies quality issues related to the data: some of the contributed entities can be assigned to wrong or implausible classes, due to individual interpretation of the submitted data, or due to misunderstanding about available classes. In this paper we propose two methods to check the integrity of VGI data with respect to hierarchical consistency and classification plausibility. These methods are based on constraint checking and machine learning methods. They can be used to check the validity of data during contribution or at a later stage for collaborative manual or automatic data correction.