This is an EarthCube GeoCODES staging ground
I edit these markdown file here
Much of this is page is on how to improve the MetaData for better:
harmonization, discoverablity, and machine-useablity
Much of what is below is about making the metadata we crawl better, but we could also consider having people author better metadata via tools like describo who’s shared-ID-lookup could someday be folded in to our jsonLD_forms-editor. (editors mentioned in other proposed tasks for 2023)
The early doc started with the main ReadMe &then topical sub-dirs
These latest (but waylaid) thoughts started in these diagrams, now below
One of the most lasting effects on the community is the use of ScienceOnSchema.org metadata (superset) DCAT3
Yet that often ends in text/string vs interconnected graph URI/thing descriptions
So I want to start improving this situation by:
deduplicating, giving URI IDs, & then using NER for annotation, to get a healthy graph of connections
By using even more of the FAIR principles, we can improve the machine action-ablility&general capabilities
The checking could also be done from the gleaner generated quads.
The dedup could even be done via openrefine, if not py lib/s
Generating an @id for (especially repeated) blank-nodes
Then lookup/find (the shared) Name (of) the entities (NER), incl linking relations
which can then be shown w/the txt, &generated from an ontology
This (iSchool) part is described in these early slides
Breakdown of present/future matching:
sparql-NB allows for cross-dataset queries, can do the dataset-page queries, & allows for more(deeper/finer):
After 1st resource-tool match will want to use other local+remote LinkedData/triples (incl eg. variable descriptions) to make better matches/kick off services
The 2nd-toolmatch-query might be where you know which type of NB/workflow you want, but need to use some local-linked(meta)data to finish off, say a webservice call in that NoteBook
There is work out there that has already gone in a similar direction:
