Crowdsourcing toponymic data in the Netherlands: Defining methods, rules and regulations for the registration of objects and names in national mapping

Abstract

For more than 200 years, geographical names have been a prominent and indispensable part of national mapping in the Netherlands. Yet, shortly after the Topographic Survey became part of Kadaster and the topographic databases and maps were given a legal basis as the Key Register of Topography (BRT), in 2009 the time-consuming field work activities that served as the main source for toponymic data were discontinued. Geographical names however, cannot be seen from aerial images and most names do not appear on or are legible from street view images either. As a result, large parts of these data could not be maintained any more, despite the legal obligation to do so. Following the principle ‘collect once, use many times’, Kadaster increasingly uses existing external data sets as a source for BRT data. Some of these contain toponymic data and are a suitable source to maintain some names categories in the BRT. This applies in particular to governmental data, such as street names and populated place names in the Key Register of Addresses and Buildings (BAG) and off-shore water names on nautical maps of the Hydrographic Service. For other names categories however, there are no sources available that meet the temporal accuracy and quality requirements for key registers like the BRT. This applies for example to many building names and area names. The past decade has shown a tremendous growth in the creation and application of Volunteered Geographic Information (VGI), collected by volunteers and free to use. Like other national mapping agencies, Kadaster already gained experience with crowdsourcing in processing user reports submitted through the BRT user feedback system, and recently, a new maintenance process was setup to check and supplement toponymic BRT data in cooperation with local historical societies. To investigate the potential of crowdsourcing as a solution to the maintenance problem of toponymic data, and as an addition to the procedures started up with local historical societies, a pilot VGI application has been built to collect the boundaries and names for area features in the BRT. Volunteers were asked to draw polygons with the approximate boundaries of geographical areas and provide the area type, the name of the area as well as their e-mail address. Thanks in part to an adequate communication, the test of the Vlakbijnamen application during a 1.5-month pilot period was a great success in terms of participation and received unanimously positive response. An analysis of the resulting data learns that the vast majority of the features seems flawless and usable, although some are missing attribute information. To avoid the registration of incorrect boundaries and erroneous or misspelled names, a dedicated team of employees with sufficient knowledge and insight in the rules and regulations of toponymic data is needed to assess the quality and implement adjustments, or to decide not to implement the contribution in the BRT. If available, cooperating local historical societies can be called in to do a last verification check. If a thorough quality assessment is guaranteed and consistent rules and regulations are provided, it is recommended to implement this method of data collection with a VGI application in the maintenance process of the BRT. Experiences from the user feedback system seem promising for the feasibility to maintain other names categories with the help of crowdsourcing as well, although the possibility to indicate the fuzzy character of area boundaries in the BRT is still to be realized. Not less important is it to finally achieve the national standardization of geographical names in the Netherlands, in order to facilitate their consistent and correct use in government and society.