One of the pillars for success for OSM is the open nature of the project. Not only can any user add data of their choice to any location on the globe but also the geometry and semantics underlay no authoritative constraints. This results in a rich and diverse dataset.
In this analyses we inquire into one aspect of the datasets diversity looking at users editing behaviour. A new methodology to investigate the tendency of users to make edits that are rare or uncommon is proposed and a thorough data analyses is conducted. Analysing the detailed mapping practices of users in OSM can be a difficult task due to the data volume and the complexity. Using the OSHDB provided by HeiGIT we estimate the rarity of a certain map action by analysing how often said action is performed.
The analyses is twofold looking at tag changes on the one hand and edits to map features types on the other hand. Rare tag edits are actions that add or delete tags that rarely occur in OSM such as unique place names or uncommon opening hours. Rare map feature type edits are actions that e.g. change the geometry of relatively rare map objects such as hospitals or cowsheds - rareness is here defined based on the general abundance of the primary tag such as highway or building. Users are then evaluated based on their tendency to contribute these rare edits to the database. We call these edits unicorn edits as they are rare but make the map a more colourful place.
Our analysis shows that contrary to our assumption there is a large number of tags and map feature types that have only few contributions. As a result, a large share of OSM users in the analysed regions exhibits a rather high tendency to map rare elements.
An index is introduced, the so called unicorn index, that allows users to compare their tendency to make rare edits to other users in the same region. The presented analyses is limited to Heidelberg and Łódź but an interactive website is made available where users can compute their personal unicorn index for Germany and Poland as well.