Usual DS uses either Python (+ Pandas, NumPy, SciPy, ...), or R to get started (or others, like SAS, Julia, ...), then to apply the results at enterprise scale, you might need either Python, Scala (or Java) There are many others solutions, but these are the most popular ones, and therefore the ones for which you'll be the more likely to find useful resources (scholarly papers, algorithms implemented in some libs/packages/, ... or even simple learning material.
From a personal experience, I really like R to get started, it's so efficient and easy to get stuff done. Then we'll face challenges to make is scale up with larger dataset (althought lots of work has been done there with things like Sparkr, Microsoft's R version, and many other things... but for enterprise pipeline scala is really nice, or Python.