In this post, I have written down some really simple instructions for setting up a working open-source data-science toolbox from scratch with anaconda. Instructions are for mac but it should be very similar for Linux.

1 - Uninstall previous anaconda (optional)

Use anaconda-clean to remove previous distribution

$ conda install anaconda-clean
$ anaconda-clean --yes

Then, remove the path ~/Users/USER/anaconda/bin from your PATH variable.

Finally, remove the anaconda folder

$ rm -rf ~/anaconda

2 - Fresh anaconda install

Download the package from here

Install it and add the binaries folder to your PATH

$ export PATH=~/anaconda/bin/:$PATH

3 - Install the necessary python and R libraries

Anaconda makes this extremely simple, installing almost every necessary library. First python

$ conda install anaconda
$ conda update conda
$ conda update anaconda

And now, R

$ conda install -c r r-essentials

Tensorflow is missing from the anaconda installation, let’s install it too

$ conda install tensorflow

aaaand this should be enough!

4 - To Do

• Discuss conda environments