Very fair question. For me, the answer is that Power BI is a nearly perfect analytical “home base”. It’s better than anything else I’ve seen at creating a seamless, intuitive user experience, while being able to integrate the best features of other analytical tools. I think the COVID dashboard that we (the JMAP team) created for Data Challenge #11 illustrates this well. In addition to the Johns Hopkins data, we pulled in probably close to a dozen other data sources, did our regression analysis and significance testing in R, our geospatial mapping analysis in Icon Map, custom visuals in Charticulator, etc. and all of that was completely invisible to the user who just saw (hopefully) an attractive, intuitive way to explore and analyze pandemic data.
Personally, I am fine with Power BI not including anything more than rudimentary statistical functions because the integration with R is so good that I would rather just take advantage of the 20,000+ custom packages available to do any sophisticated analysis I could imagine, and then just pull those results easily into Power BI.
Similarly, if you look at some of the amazing machine learning analyses that Enterprise DNA Expert @bradsmith is doing, he is also using Power BI extremely effectively in this “integrator” role, incorporating the best machine learning algorithms and capabilities from Python and R together in a seamless report.
However, I don’t work with datasets with nearly as many fields as you do, and so the size and complexity of your data ultimately may be better suited to some of the tools you mentioned. With regard to business versus academia, I do think it’s true that Power BI is marketed most directly to business users and I have no idea what level of traction it’s gaining within academia. However, I do know that eDNA members @Tanzeel and @dsirias are university professors using Power BI extensively, so they may be able to offer insight into that question.
I hope this is helpful.