Quickstart/Installation

Install using pip

You can install Datamode with pip from your shell/command-prompt:

pip install -U datamode
jupyter notebook

Then from your Jupyter Notebook:

from datamode.interface import *
quickshow('https://raw.githubusercontent.com/datamode/datasets/master/movies.csv', sample_ratio=1)

Play around with the interface a little. Then change the code block to this:

from datamode.interface import *

tcon = run_transforms([
  SourceFile('https://raw.githubusercontent.com/datamode/datasets/master/movies.csv', sample_ratio=1),
  CanonicalizeDate('release_date'),
  Expression('roi = revenue / budget'),
])

See what happens when you click on the histograms or the transforms list. Also, try to delete or change some of the above transforms, or change sample_ratio.

You’re ready to try our howto guides now: How To Guide.

Note

Datamode currently requires Python 3.6+ (64-bit). See below if you have any problems installing.

Install using conda

If you don’t have a conda virtualenv, follow these steps:

# Create a conda env and enter it - then you can use pip.
# Follow the instructions above to run Datamode in Jupyter.
conda create -n dm_env python=3.7
source activate dm_env
pip install datamode
jupyter notebook

If you already have a conda virtualenv, make sure it’s using Python 3.6+, because some conda installs run Python 3.5. You can do import sys; print(sys.version) to find out what version you’re using.

Create a virtualenv

If you aren’t already running Python 3.6+, you can create a virtualenv to host Datamode. The easiest way to do that is to install virtualenvwrapper:

pip install virtualenv virtualenvwrapper
mkvirtualenv -p python3 dm_env
workon dm_env

If you’re on windows, install virtualenvwrapper-win instead:

pip install virtualenv virtualenvwrapper-win
mkvirtualenv dm_env
workon dm_env

You’ll need to re-enter the virtualenv with workon when you start a new shell/command-prompt.

Python architecture (64-bit vs 32-bit)

Datamode only supports 64-bit Python, which you’re probably already using. Some versions of conda use 32-bit Python. If you want to make sure you’re using 64-bit, you can run this code in Jupyter or any Python environment:

import sys
print('64bit' if sys.maxsize > 2**32 else '32bit')