Both -- it's mostly about the strong, growing community, but that derives from Python being a versatile language. Python is relatively easy to learn and use, can be quick and dirty, fast enough for most applications*, and has a ton of quality libraries, making it a great ecosystem for ML research.
It's worth noting that most ML researchers come out of academic labs where solid SW engineering isn't required, so Python is an obvious choice (over e.g. C++) for ML researchers in academia and industry. As a result we see main ML libraries and open-source projects written in Python -- TensorFlow, Keras, scikit-learn, etc. -- which further proliferates the adoption and use of Python for ML.
*After prototyping and optimizing an algorithm in Python, we may implement it in C++ for the speedups.
We're a place where coders share, stay up-to-date and grow their careers.
We strive for transparency and don't collect excess data.