Wisdom about the Crowd: Assuring Geospatial Data Quality Collected in Location-Based Games

The idea of outsourcing geospatial data creation tasks to the crowd (volunteered geographic information, VGI) has become quite popular in the field of geographic information science (GIScience). As one approach to VGI, location-based games (LBGs) have been shown to be successful in motivating non-expert users to collect and tag geospatial data. However, the central VGI problem of data quality has so far not been solved satisfyingly. Previous studies show that games that reward their players for inor post-game data reviewing can assure only a validation rate of about 40% of the data. We address this problem with a new LBG design pattern, based on game rules that encourage players to take the decisions of others into account while making their data collecting decisions. We empirically evaluate the new pattern by comparing the positional accuracy of data collected with two different rule sets for the LBG GeoSnake. Our pattern is shown to result in a significant accuracy improvement.