Keeping Computational Fluid Dynamics (particularly using OpenFOAM) post-processing in Python allows significant advantages, to my mind. Transparency (did I hack my stats models to get lovely pictures that fit the data then say can't show it because of IP?) & risk reduction (does access to foundational code go missing when some business owner gets pissed-off?), namely.
Specifically, access to a whole ecosystem of data analysis capabilities [1]: the ability to hook into R-based [2] statistical modules from Python [3], in addition to Python-based modules [4]. Can be visualised in-depth via PyVista [5].
Other's thoughts? Alternative hypotheses very welcomed!
[1] https://jakevdp.github.io/PythonDataScienceHandbook/
[2] https://www.r-project.org/
[3] https://rpy2.github.io/
[4] https://www.statsmodels.org/stable/index.html
[5] https://tutorial.pyvista.org/index.html