Hi, @didibus, interesting, what a thorough list!!
See also this related list.
It would be a great idea to keep such lists in mind while going into use cases (e.g., translating existing Kaggle kernels to clojure), and seeing which kinds of functionality are actually important.
Regarding visualization, the libraries you mentioned have huge potential. Still, to get an idea what is missing, you may look into galleries like these:
htmlwidgetss | ggplot2 | R in general
and think, which of these visual elements are already easy to implement, and which are not.
Regarding scikit-learn, it is worth watching fastmath and tech.ml (under construction, growing fast).
Regarding Numpy, Scipy, it is worth looking into the discussion in this Zulip topic.