GraphLab Create Quick Start

Need help with a project now? The GraphLab team can help.

Learn to code machine learning apps.

Have a question?

Already have GraphLab Create and want to get the latest version? Follow the upgrade instructions.

1
Get Your Product Key

Join the beta! It is free and easy. Sign up and instantly receive a GraphLab Create product key for your individual use. We will also send you a confirmation email.

All fields are required.
By signing up, I agree to the GraphLab Create EULA.

Welcome to the beta! Your product key has been generated. To make configuration easy, we have provided a shell command that will insert your product key in a GraphLab Create configuration file located in your home directory. Before running Python, paste the following code into your terminal window and execute. GraphLab Create will reference this file upon import.

(mkdir -p ~/.graphlab && echo -e "[Product]\nproduct_key={key}" > ~/.graphlab/config && echo "Configuration file written") || echo "Configuration file not written"

Successful completion of this step will show "Configuration file written" in your terminal window. Now, proceed to Step 2.

A product key has already been generated for this email address. Would you like us to resend your confirmation email to view your product key?
The confirmation email has been sent to your registered email address. Check your email to view your product key. Follow the link in the email to return to this Quick Start and proceed with the installation.
2
Install GraphLab Create (beta)

GraphLab Create is easy to install is easy to install.

Supported operating systems:


What do I need to do?

  1. Ensure your system is running Python 2.7.x (execute python -V in terminal)
  2. Install pip, a Python tool for installing Python packages
  3. Choose the below installation method that works best for you

You can install Graphlab Create system-wide (recommended) or in a Python virtual environment (virtualenv).

Copy and paste the following code into your terminal window and hit "Enter".

sudo pip install graphlab-create==0.9.1

If you are unsure about installing or upgrading these libraries system-wide we recommend installation with virtualenv.

Installing with in a virtualenv contains the installation of GraphLab Create and allows you to customize this virtualenv with other Python packages you may want for your data science projects.

  1. Ensure your system has virtualenv installed. To verify, execute pip freeze. To install, execute sudo pip install virtualenv in your terminal before proceeding
  2. Copy and execute the following commands in your terminal. This will create a virtual environment called 'graphlab' and install graphlab create version 0.9.1
virtualenv graphlab
. graphlab/bin/activate
pip install graphlab-create==0.9.1
Resource: Learn how to use virtualenv


Detailed installation instructions for Mac OS X and Linux

3

Start Creating: Hello World

Now you can start using GraphLab Create. Let's build a recommender. Copy and paste the following code into your Python console.

import graphlab as gl
url = 'http://s3.amazonaws.com/GraphLab-Datasets/movie_ratings/training_data.csv'
data = gl.SFrame.read_csv(url, column_type_hints={"rating":int})
data.show()
model = gl.recommender.create(data, user_id="user", item_id="movie", target="rating")
results = model.recommend(users=None, k=5)

You've just used the fundamentals of GraphLab Create! To learn more about his recommender see this notebook.

What's next? Go learn with GraphLab Notebooks!

The Getting Started with GraphLab Create is a good notebook for beginners. Download the code and have fun. It will give you a broad overview of how to use GraphLab Create. We’ll introduce the SFrame, a tabular structure ideal for data manipulation and feature building, and the Graph which is a structure ideal for sparse relationship data. We’ll also introduce our machine learning toolkits. Using these, you’ll ingest data, build a graph, and run a graphical model to generate an insight. Check out all of our notebooks!, there's a lot to see.

Need help with a project now? The GraphLab team can help.

Learn to code machine learning apps.

Have a question?