I am a newbie to OpenStreetMap and was wondering what the preferred method for mapping shops is? I am a bit confused as many people mark out the buildings then place a single node for the shop inside the building. However many people also mark the building area itself as a shop. If it is a large building with smaller shops within one building I guess the building should be mapped and nodes added for the shops inside but what if it is for a shop that occupies the whole building? What is the preferred method? I think this is confusing as some shops have a relatively small area where customers go with the larger part of the building for storage etc. In this case should building:part be used? Also, I see many shops where they are all part of what was the same building but dividing walls have now been put in and extensions added at the back. What is the best approach for mapping this case? Sorry to put so much in one question and maybe I just need to spend more time on the wiki. |
As with many other issues in OSM, you have to be aware that mapping practice varies quite a bit. That being said, I think there is generally a consensus about the following statements:
The remaining details are less clear: For single-use buildings, many people add the shop/amenity tags to the building itself. However, this expresses two features, the building and the shop/amenity within, as only one OSM element. and thus fails when they e.g. have different names. Therefore you may choose to map the shop/amenity with its own node or area even in this situation. For buildings containing multiple POI, you generally have the option of either adding multiple nodes – this is the most frequently used option as indoor maps based on OSM are still only experimental – or smaller areas covering only the sections/rooms of that particular shop/amenity. It is usually not necessary to also tag such smaller areas as building parts, unless it makes sense to do so for other reasons (e.g. based on the externally visible building structure). answered 03 Jun '13, 14:44 Tordanik |
With more precise routing nowadays, I prefer tagging every POIs (shops, etc.) as nodes near their respective entrance (entrance=shop or entrance=main) on the building boundary. When you use a routing engine to find directions to go to a specific POI, it will then route to the node itself or the nearest entrance, which is more precise than the default building centroid, which is sometimes several meters away from the entrance. Using only nodes over buildings also makes it easier to update/add/remove shops later on without loading the building geometries. Finally, when someone wants to export POIs as JSON or GeoJson with overpass, it will make the files way smaller, because it does not need to include all building geometries. answered 07 Jun '22, 19:03 ChaireMobili... |