Setup EC2

Create instance via awscli or console. P2 instances using Deep Learning AMI are preferrable for simplicity. Be sure to configure security group and SSH keys to restrict access to IP(s).

Access

Follow the ssh login directions for the instance and activate one of the default venv configurations:

$ source activate tensorflow_p36

Setup Jupyter

To access an IDE, one needs to setup the server first. Then configure the environment:

$ pip install jupyter_contrib_nbextensions
$ jupyter contrib nbextension install
$ pip install jupyterthemes
$ jt --theme grade3 -fs 9 -nfs 11 -tfs 11 -cellw 90% -T

Install Tools

For this notebook, library dependencies are short, so no requirements file:

$ pip install --upgrade numpy pandas scipy featuretools xgboost seaborn scikit-learn tables pyarrow h2o

Setup GH Connection

Create keys and add them:

$ ssh-keygen -t rsa -b 4096 -C "your_email@example.com"
$ ssh-add -K ~/.ssh/id_rsa
$ eval "$(ssh-agent -s)"

Got to GH key settings and add.