I'm an expert at 1 and 4 and know all 6. Mathematica is by far the most powerful and most complete. In fact Jupyter notebook (then lab) were developed to copy Mathematica but that effort fell far short. Arbitrary precision and tracked precision calculations are automatic in Mathematica so you don't have to worry about roundoff if you don't want to. Functions, including statistical go way beyond what the others have in breadth and depth even if you import multiple libraries with the others. Derived distributions is a simple command for even complex functions, albeit they may need to be evaluated numerically for many situations. But at least they have a symbolic representation which can be used anywhere. The easy mix between symbolic and numeric, even within a single calculation provides power unmatched by any of the other options. Every field of Math is available at your fingertips and much, much more than math. For example, the data that is available with simple commands is astounding and again unmatched. For example, off the cuff, try the following for Countries:

Graphics[

Table[ {Hue[ Random[]],

Tooltip[ CountryData[ whatCountry, "Polygon" ],

Style[

Row[ {whatCountry, "; Population = ",

CountryData[ whatCountry, "Population" ] / 10.^6,

" million"}], 24, Red ]]},

{whatCountry, CountryData[ "MiddleEast" ]} ]

]

Or

DateListPlot[{CountryData[

"UnitedStates", {{"NationalIncome", "Local", 2000}, {2000,

2020}}],

CountryData[

"China", {{"NationalIncome", "USDollars", 2000}, {2000, 2020}}]}]

And the language is so concise. Even Python seems very verbose by comparison. Of course, this comes at the price of a steep learning curve but the "unlimited power" (to channel Emperor Palpatine) ...