Assessment of the accuracy of GeoNames gazetteer data

Abstract

Gazetteers are the basis of many geospatial applications and serve an important role to collect and make available knowledge about the physical world such as place names and their coordinates. GeoNames is one of the largest and most often used gazetteer and it is generally assumed to be of sufficient quality. In this paper, we examine the quality and accuracy of the data in more detail, triggered by some anomalies encountered during its use. We present a classification of inaccuracies ranging from grid patterns, imprecise coordinates, overlaps and repetitions as well as misclassifications and visualize these for a range of countries. We finally give an outlook into potential corrections.