I have been using R and RStudio for prototyping and model building and due to some persisting problems (which would only be applicable to the environment that I am using in) we have decided to use python. I am looking to know:

  1. is there a python development environment like "RStudio Server"?
  2. how easy/hard to enable multi-threading/multi-core processing in Python for decision trees/gradient boosting?
  3. can data persist in python dev environment server? Meaning: Can I save the dev environment with code and data frames and come back the next day to access/pick up where I left?

  4. Can you connect to Oracle database from Python?

  5. Is there a way I can import .RData to python development environment?

Why ask these questions here?

While looking up on search engines, I am mostly getting unreliable results, training insitutes promotions, outdated blogs and whitepapers from industry giants. I need a reliable answer.

  • $\begingroup$ 2): Algos in sklearn and also the interfaces to XGBoost and lightGBM are as parallelized as their analogons in R. $\endgroup$ – Michael M Apr 4 '18 at 16:00

1) Anaconda Spyder, maybe

2) sklearn, random forest has an option to select the number of jobs, and it will take care of parallelizing

3) i don't think so, but you can pickle objects and load them up. you can probably do something like create a variable_name - value dictionary and just pickle that

4) never tried

5) according to https://stackoverflow.com/questions/21288133/loading-rdata-files-into-python, I don't think so

  • 2
    $\begingroup$ Why pickling when you have feather format? $\endgroup$ – Aditya Apr 5 '18 at 4:43

I can try to answer the 3rd question. You can use Jupyter + Python Kernel (install Anaconda to use conda commands with python Kernel). Jupyter allows keep your code and notes and graphs in one notebook and save it.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.