As more and more companies such as Mapbox, Telenav, and CartoDB begin to rely on OpenStreetMap (OSM) in their business models, it’s important to understand whether OSM is achieving comprehensive coverage across both urban and rural landscapes. Small cities thus become an ideal laboratory for studying the “crowd” behind OSM because (a) they help us understand whether OSM is achieving coverage beyond major urban centers, and (b) the smaller corpus of edits and contributors allows the researcher to better comprehend what is occurring in the place. In short, when we look at OSM in any given place, whose influence do we see? And what brought them to the place if they do not live there?
Small multiple maps are one way to visualize the full set of contributors and the influence of each. This map shows all contributors to OSM in Tres Arroyos, Argentina between the project’s inception and the end of 2014. The maps are sorted from top to bottom (reading left-to-right) based on the contributor’s number of unique days active in the project. Maps such as these can highlight the division of labor between the more casual mappers, the more systematic mappers, and automated activities such as bot fixes or imports.
When studying the free-form comments left by OSM contributors and linking it to the geographic data , it becomes clear that small cities in OSM are a crossroads of hobbyists, bots, paid mappers, imagery tracers, importers, and a small handful of people who know and love the region, but may not be from the town itself. This broad variety of influences disrupts the original notion of volunteered geographic information as being contributed by individual “citizen sensors” sharing primarily local knowledge.
View a talk about the project
This talk at State of the Map US 2015 focuses on OSM construction in the city of Salina, Kansas. It compares growth in this town with larger cities in the region and other small towns in North and South America.
Quinn, S. Using small cities to understand the crowd behind OpenStreetMap. GeoJournal. doi: 10.1007/s10708-015-9695-6 [PDF (Author’s accepted version)]