One of the problems I quickly came across while working on Kaggle competions was the issue of computing power. Often I'd want to run an algorithm on a large dataset, perhaps cross-fold validation or the like, and find that training the training time took FOREVER!
This frustrated me to no end. I'm learning Python and Jupyter now, and one of the initial things that struck me was that its use is almost ubiquitous on Linux. Now, I'm a long time fan of all things Linux, and I definitely am no stranger to to Oracle's Virtualbox - but I very quickly realized that using a virtual machine (on my laptop) would not quite live up to the challenge of meeting my needs - but lo and behold, there is an answer!
Enter the remote server.
Now we live in an age of the cloud, where recruiting additional computing power is nothing more than a few mouse clicks away - so I decided to look at my options.
Digital Ocean is a great, great way for beginners to get their hands dirty. Setup is quick, simple, and best of all, cheap. A great guide to get you up and running an be found here:
https://data36.com/data-coding-101-install-python-sql-r-bash/
If you access the site through that guide, you'll also get a sweet 10 credit! Very nice.
Now I'm new with servers, so the second problem I came across very quickly was in transferring Data between the client and the server. How do we make that happen? No problem: We have filezilla for that
Filezilla provides a GUI for file management and transfer between client and server, making transferring large datasets between client and server a breeze. I included a quick and dirty guide that should get you up to speed right here:
http://datascienceguide.github.io/beginner-tutorial-how-to-get-started-with-data-science-using-servers
I have to say - my explorations into Python and Data Science are proving to be rather interesting. Ideally, I want to very quickly get through the basics, acquainting myself with the way Python implements common algorithms. Then after that, I want to start writing scripts, and taking the plunge into deep learning.
We'll see where this goes!