Ask HN: Machine Learning: Easiest way to transition from R to Python

5 points·uptownfunk··5 comments

I'm a data scientist, want to broaden my skills and expertise and also learn python/numpy/scipy/matplotlib/sklearn/etc.... What's the closest python ide similar to R/RStudio?

View on HN

I'm not familiar with R, but on the Python side, aside from the already-mentioned excellent Jupyter Notebook, you might also be interested in checking out Spyder, which is more of a dedicated IDE. Download/screenshots: https://github.com/spyder-ide/spyder Docs: http://pythonhosted.org/spyder/

EDIT: If you're on Windows, check out the Anaconda python distribution. I'm told it makes installing all the libraries you mentioned as easy as it is on *nix systems: https://www.continuum.io/downloads

Not an expert here, but I have dabbled with both R and Python.

For learning and even publication purposes, I'd consider looking at Jupyter Notebooks: http://jupyter.org/

There's still some configuration involved to get it to work, but most of the required libraries should come bundled in the installation.

As for learning those tools -- probably just do a few quick tutorials to get the gist of Pandas/Numpy, and then slowly try migrating small data projects that you'd normally do in R to Python.

Take a look at Rodeo from yhat.com.

The screen layout is very similar to RStudio and you can see the inside of variables and data frames like in RStudio.

Btw - Yhat developers also hang out here in HN.

how is package installation?

I use conda because I'm a bit control freak; but you can search and install packages directly from Rodeo. There's a tab for that ;-).

Reply on news.ycombinator.com