Intrinsic assessment of the temporal accuracy, up-to-dateness, lineage and thematic accuracy of OpenStreetMap

Sunday 12:00, S.1.3

Francesco Frassinelli ¹, Marco Minghini ¹, Maria Antonia Brovelli ¹ 30 minutes

¹ Politecnico di Milano, Department of Civil and Environmental Engineering, Milano, Italy

A well-known factor which still limits the widespread use of OpenStreetMap (OSM) data is the concern about its quality, which is measured through a variety of parameters such as positional accuracy, completeness, thematic accuracy, temporal accuracy, up-to-dateness and lineage. To date, many approaches have been proposed to assess these parameters for OSM. These approaches can be classified into two categories: extrinsic, when quality assessment happens through the comparison against an external reference dataset (e.g. a governmental map), and intrinsic, when quality assessment happens through the sole analysis of OSM itself. The present work compensates for the almost total absence of intrinsic approaches to assess the OSM quality parameters of temporal accuracy, up-to-dateness, lineage and thematic accuracy. Based on the OSM Full History Planet File and a stack of open source components, a web-based tool was developed (https://is-osm-uptodate.frafra.eu) which, for any selected area, offers thematic visualizations of OSM nodes and ways according to: the date of creation (i.e. first edit); the date of last edit; the number of versions (i.e. the total number of edits); the update frequency; and the number of different contributors who have edited that node or way. A spatial analysis of the results allows to draw some general conclusions. First, temporal accuracy, up-to-dateness and lineage of OSM data are usually higher in urban areas (more “appealing” to map and where more contributors are active) than rural areas. This is in agreement with the results usually found in literature for OSM completeness. However, although the global trend is quite well-defined, at the local level OSM shows very heterogeneous patterns due to the actual presence and mapping activity of contributors, which can be very different even between neighbouring towns. An additional analysis was focused on OSM thematic accuracy, i.e. the accuracy of tags associated to OSM nodes and ways, and its temporal trend for specific object categories (commercial activities, street furniture, natural elements, etc.). For these categories, despite again the typical heterogeneity of OSM contributions, some common recurrences were detected in the order in which tags are added; consequently some trends were defined which, based on the tags available and the order and time of their appearance, can automatically indicate their thematic accuracy.