Biology and biomedical research is the main advanced application I am aware of. Its important enough a set of domains to show there is value in semantic technologies (rdf, owl etc)

I have been wondering why it got adopted there and not elsewhere. Its not clear. Adoption or not of software and information tools is a mix of cultural and economic reasons beyond pure technical aspects.

CSV, json etc are information black holes. Eventually people will realize that the systematic annotation of data within a metadata framework is not fanciful but an enabler of more confident and robust information exchange and processing

[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4652617/

The key word in my comment is "integrate", not simply producing data. We're using it for language data in my field, e.g. currently I'm integrating the Danish WordNet with the English one + various other datasets.

If you check out the common ontologies in use you will get an idea which areas are covered: https://lov.linkeddata.es/dataset/lov/