fils
commented
3 years ago @aorehek so I don't know what the current thinking is on SOS as a group. So please note these are just my personal observations.
FYI.. check https://github.com/ESIPFed/science-on-schema.org/issues/105 for some ideas on this question too.
Personally, my POV is that schema.org spatial is of little use unless you want to convey it to Google Dataset Search. I suspect most indexers and users of spatial data downstream will find shema.org spatial too burdensome to support (I did) or will do so at an expense likely greater than its value.
Eventually you work your way down something like https://schema.org/GeoShape to define geometries with, a 9 levels of schema.org spatial journey.
My own (not group) recommendation is to use geosparql. https://github.com/opengeospatial/ogc-geosparql
There are other approaches (some shown in the issue above) and groups like https://github.com/opengeospatial/SELFIE have taken other approaches which I think are just as valid. Indeed, your possible approaches are many and you likely would go with one best connected to your functional needs and architecture.
I'm not a spatial person (notice how I didn't mention that till the end) ;) , so I'd enjoy hearing back from you if any of these approaches works for you and where you went with this. Hope this helped..
What would be a good way to describe a spatial type, like "raster" or "grid" or "point"? The closest thing I can find on Schema.org is under Geospatial Geometry: https://schema.org/GeospatialGeometry. Even there, I can't determine what is the best way to convey this information. In the data I'm working with, this is an important metadata because a majority of the datasets contain this information.