Need support with conversion of a correspondence table to rdf (an example of direct correspondence using skos vocab)

[agneta20]

I tried to develop a direct correspondence between YSTAFDB and Exiobase flows and activities using skos vocabulary. I chose to do it only for phosphorus flows (as these flows are the only ones available for 2011 which the data in Exiobase corresponds to). I have made a selected choice of flows which can be broadly corresponded to the flows and activities in the supply and use tables (which is the only part of exiobase converted to RDF currently).

correspondence_Phosphorusflows.xlsx

[mfastudillo]

Hi

we do not have yet a general procedure, but I think it would make sense to go from a (csv) table with 3 columns

subject	predicate	object
phosphate rock mining (PRM) refining	skos:narrower	Mining of chemical ...
...	...	...

and transform the labels into their corresponding URIs. This "transformed" table can be combined with a (datapackage) file of metadata of the table (provenance or whatever)... and then have function in the library that creates a turtle file based on both. I am not sure why there are more than 3 columns in the correspondence example tab.

[agneta20]

In YSTAFDB, the flow data is given with the activity types it is related to. So each row is a flow that consists of the activity the flow is an outputOf and inputOf. Hence there are more than 3 columns. If I turn this into the three column data forum, i think the information is not the same.

[boweidema]

If the flow-object names in YSTAFDB are not unique, i.e. if the same name is used for different flow-objects depending on which activity it comes from, then this should be fixed before converting to RDF. e.g. by creating "flow-object from activity". However, this should probably best be done in cooperation with the YSTAFDB folks.