The term "Semantic Web" goes back to at least 1999 and the idea – enable machines to "understand" information they process and enable them to be more useful – is much older. But still we do not have universal expert systems, despite that they would be very advantageous, especially in the context of (interdisciplinary) research and teaching.
My impression is, that from the beginning semantic web technologies was dominated by Java-based tools and libraries and that the situation barely changed until today (2022): E.g. most of practical ontology development/usage seems to happen inside Protegé or by using OWLAPI.
However, in the same time span we have seen a tremendous success of numerical AI (often called "machine learning") technologies and here we see a much greater diversity of involved languages and frameworks. Also, the numerical AI community has grown significantly in the last decade.
I think, to a large part this is, because it is simple to getting started with those technologies and Python (-Interfaces) and Jupyter-Notebook contribute significantly to this. Also, Python greatly simplifies programming (calling a library function and piping results to the next) for people who are not programmers by training such as physicists, engineers etc.
On the other hand getting started with semantic technologies is (in comparison) much harder: E.g. a lot of (seemingly) outdated documentation and the lack of user-friendly tools to achieve quick motivating results must be overcome in this process.
Therefore, so my thesis, having an ecosystem of low-threshold Python-based tools available could help to unleash the vast potential of semantic technologies. It would help to grow the semantics community and to enable more people to contribute contents such as (patches to) domain-ontologies, sophiticated queries and innovative applications e.g. combining Wikidata and SymPy.
Over the past months I collected a number of semantic-related Python projects, see https://github.com/pysemtec/semantic-python-overview. So the good news is: There is something to use and grow. However, the amount of collaboration between those projects seems to be low and something like a semantic-python-community (i.e. people who are interested in both semantic technologies and python programming) is still missing.
Because complaining alone rarely leads to improvement of the situation, I try to spawn such a community, see https://pysemtec.org.
What do you think, can Python help to generate more useful applications of semantic technology, especially in science? What has to happen for this? What are possible counter-arguments?