Mapping complex sites

I’m trying to improve the mapping for the Denmark Hill campus of King’s College London, where I work. As you can from the KCL maps the sites are complex, as the University shares the area with two hospitals (e.g. the intermingled yellow and purple buildings in the second map).
Currently, the sites are mapped using several separate areas, rather awkwardly drawn to include one set of buildings and omit others. I think a better method would be to create a site relation and add the buildings as members. I would need to add a few nodes to position the labels for the site. The only down side is that there would not be a colour-coded area on the map, but given two intermingled sites I’m not sure this is appropriate.
Any feedback would be welcome, before I go and make a bunch of incorrect changes.

The colors are just a matter of writing the correct rules for a renderer. When you deploy one yourself, you can do this (I have no experience with this).

I would not “add a few nodes to position the labels”. Nodes need to represent something, a building, a gate, etc. It’s better to properly add the physical objects of the campus (buildings, gates, cafeterias, libraries, etc.) as nodes or polygons. Those nodes and polygons can have a name tag. A node with only a name tag is suspicious to me (aka mapping for the renderer) and should be avoided.

Hope this helps a bit. Enjoy mapping the campus

I think site relations are a good way to go on this.

Regarding adding “a few nodes to position the labels”, don’t add those nodes to OSM. However I think I know where you are coming from: For creating a printed map of an area I volunteer at I ended up creating some nodes and ways to position labels. But I kept those nodes and ways in a separate non-OSM database and simply set up the rendering software to use data from both the OSM extract and my private local data. Maybe something like that would work for you too.

Nodes do not always represent something physical on the ground - for instance with boundaries. And it can be compared with doing other types of drawing, particularly CAD, then construction lines are often useful. I’ve used some before for a geometric area on OSM and left them on the map till I’d completed it. On paper maps that is to some extent what latitude and longitude lines are. If there was some tag to explain what they are and that renderers should ignore them, I don’t see a problem.

I’m aware that there are nodes that are tagged place=city. However, you should not tag a node with only e.g. name=H just to have and “H” on the map Then it’s better to map the building and tag it with building=yes, ref= H

Yes, I was thinking about the place=* key, although looking at the wiki page, it doesn’t really have anything for a site. I could just try creating the relation and see if it is labelled on the map according to its name on

but the site is a college, right ? so the relation will be tagged type=site, amenity=college, name= Denmark Hill campus
But I don’t know whether the standard map style recognizes the site-relation.