What we dislike about Excel is summed up in the memes. "No, that number is not a date" problems are just inexcusable.

But in terms of computing power, it's awesome.

My personal preference is a sweet spot to use Python with xlsxwriter or openpyxl to read (and write) worksheets where I can't do it inside Excel due to my lack of competency. Then I can produce Excel charts (even directed by the Python) which are indistinguishable from what others do, in style terms. I guess if somebody writes a graphics output engine which is 1:1 congruent with data uplift to Powepoint, I'd stop.

(the thing I find hardest is to select cell contents into a new array, not to compute a map/reduce over them but simply to populate a table in a new specified location as an XFILTER over some other nominated cells. COUNTIF is fine. But thats a map/reduce solution. I want the literal subset under some criteria, so I can do other things with it, and I want to see it not just have it a computed state off-workbook. The evidence is strong you have to write VBSCRIPT solutions to iterate, and none of them are well documented that I can find)