Is Mathworks losing out on data science applications?

Dear Mathworks team,

I am a great fan of Matlab, having used it for many years and maybe even considering myself an expert. In the last two years though I am feeling more and more pressure from data science applications to move to Python. With places like Kaggle massively promoting it and new software packages coming out which often have a Python connector, I ask myself: Is Mathworks losing its ground in data science? Honestly I love the Matlab IDE and the debugging capabilities which come along with it. But I don't feel Mathworks is taking people by the hand with regard to libraries in other languages. With all the open source libraries available, doesn't Mathworks recode popular ones up to be used in Matlab? Some concrete examples:

• ScikitLearn in Python is very popular. Did Mathworks close the capability gap? Are study methods named differently?
• I have seen Mathworks working together with Kaggle. Mathworks on the other hand should have noticed that a lot of competitions were won with the XGBoost library. There are several connectors available...but where is the Matlab one?
• Tensorflow by Google has skyrocketed in popularity since its introduction. Where is the Matlab connector?

I understand that it may be more profitable in the short term to stick to a closed environment or let the community do some work. But I have the faint feeling that this strategy does not go well moving forward with new generations being more and more focused on environments such as Python. I would be very glad if Mathworks would go to some lenght and close some gaps. Glad to hear what you think.

Or did I misinterpret things and am supposed to use something like described here: https://uk.mathworks.com/help/matlab/examples/call-python-from-matlab.html#responsive_offcanvas

Best regards,

Florian