> For many data science projects, maintaining large code bases in Python is often the optimal decision (rather than reinventing the wheel and writing your own d

> For many data science projects, maintaining large code bases in Python is often the optimal decision (rather than reinventing the wheel and writing your own data frame and machine learning libraries, or using immature poorly maintained ones in other languages — for data manipulation and scientific algorithms, these libraries are highly optimized in Python anyway since the underlying code is in C or Fortran) Or you could use R, instead of the half-baked immature clones in Python :) I'm only slightly joking here, even if Python is much much better for string processing, it's much less useful for DS tasks (unless it's NLP or DL, to be fair). That being said, Python is the second best DS language, as well as second best for everything else, so it can be a good choice for these projects.