Hi Hendrik, I actually thought about this when making the article. My goal here wasn’t to compare the performance of Python packages that likely have operations written in C. While there certainly could be a performance gain, I don’t think it would shave off that much time.

I wanted to compare straight Python to straight Julia — for those that write low-level algorithms but want to remain high-level. So with that in mind, the Linear Regressor is just one example of very many iterative tests.

I don’t think the speed issues really came from these functions as well, because too be honest they are exactly what one might expect with Python. It seemed like Python was just as good at Julia for quick operations, but when it came to working with huge arrays of data, Python just wasn’t happy.

I certainly see your point. If you would like to try it out I’m sure it would be fairly easy to replace the two functions in my notebook with Numpy functions. Perhaps whenever I am at the computer I will try it real quick, as well, and see what kind of performance benefit we may get.

Thank you for your comment!