Set up the tutorial data and initialize a Data Context

Welcome to the Great Expectations Getting Started tutorial! This tutorial will walk you throw a simple exercise where you create an Expectation suite that catches a data issue in a sample data set we provide.

Prerequisites for the tutorial:

  • Make sure you have Docker installed on your machine.

  • Have git installed on your machine to clone the tutorial repo.

  • You need a Python environment where you can install Great Expectations and other dependencies, e.g. a virtual environment.

About the example data

The NYC taxi data we’re going to use in this tutorial is an open data set which is updated every month. Each record in the data corresponds to one taxi ride and contains information such as the pick up and drop-off location, the payment amount, and the number of passengers, among others.

In this tutorial, we provide two tables, each with a 10,000 row sample of the Yellow Taxi Trip Records set:

  • yellow_tripdata_sample_2019_01: a sample of the January 2019 taxi data

  • yellow_tripdata_staging: a sample of the February 2019 taxi data, loaded to a “staging” table so we can validate it before promoting it to a permanent table

If we compare the passenger_count column in the January and February data, we find a significant difference: The February data contains a large proportion of rides with 0 passengers, which seems unexpected:


The data problem we’re solving in this tutorial:

In this tutorial, we will be creating an Expectation Suite for this example data set that allows us to assert that we expect at least 1 passenger per taxi ride based on what we see in the January 2019 data. We will then use that Expectation Suite to catch data quality issues in the staging table, which contains a new batch of data every month.

Set up your machine for the tutorial

For this tutorial, we will use a simplified version of the NYC taxi ride data. We prepared a Docker image that contains a local Postgres database with the data pre-loaded, so you can easily get up and running without any local dependencies.

Clone the ge_tutorials repository and start up the container with the Postgres database containing the data:

git clone
cd ge_tutorials/ge_getting_started_tutorial
docker-compose up --detach

You will now have a Postgres database running in the background with the pre-loaded taxi data! In case you want to connect to the database, to check out the data, you’ll find instructions in the README in the repository. But for now, let’s move on!

Install Great Expectations and dependencies

If you haven’t already, install Great Expectations. The command to install is especially great if you’re a Dickens fan:

pip install great_expectations

You’ll also need to install a few other dependencies to connect to your Postgres database. In general, Great Expectations doesn’t clutter up your deployments by installing dependencies before you need them.

pip install SQLAlchemy psycopg2-binary

Create a Data Context

In Great Expectations, your Data Context manages your project configuration, so let’s go and create a Data Context for our tutorial project!

First, we want to create a separate project directory ge_example/ for our tutorial project. The ge_tutorials repo contains the final version of this tutorial, but we want to start from scratch here, so let’s move out of ge_tutorials:

cd ../..
mkdir ge_example
cd ge_example

When you installed Great Expectations, you also installed the Great Expectations command line interface (CLI). It provides helpful utilities for deploying and configuring Data Contexts, plus a few other convenience methods.

To initialize your Great Expectations deployment for the project, run this command in the terminal from the ge_example/ directory:

great_expectations init

You should see this:

  ___              _     ___                  _        _   _
 / __|_ _ ___ __ _| |_  | __|_ ___ __  ___ __| |_ __ _| |_(_)___ _ _  ___
| (_ | '_/ -_) _` |  _| | _|\ \ / '_ \/ -_) _|  _/ _` |  _| / _ \ ' \(_-<
 \___|_| \___\__,_|\__| |___/_\_\ .__/\___\__|\__\__,_|\__|_\___/_||_/__/
             ~ Always know what to expect from your data ~

Let's configure a new Data Context.

First, Great Expectations will create a new directory:

    |-- great_expectations.yml
    |-- expectations
    |-- checkpoints
    |-- notebooks
    |-- plugins
    |-- .gitignore
    |-- uncommitted
        |-- config_variables.yml
        |-- documentation
        |-- validations

OK to proceed? [Y/n]:

Let’s pause there for a moment and take a look under the hood.

The great_expectations/ directory structure

After running the init command, your great_expectations/ directory will contain all of the important components of a local Great Expectations deployment. This is what the directory structure looks like:

  • great_expectations.yml contains the main configuration of your deployment.

  • The expectations/ directory stores all your Expectations as JSON files. If you want to store them somewhere else, you can change that later.

  • The notebooks/ directory is for helper notebooks to interact with Great Expectations.

  • The plugins/ directory holds code for any custom plugins you develop as part of your deployment.

  • The uncommitted/ directory contains files that shouldn’t live in version control. It has a .gitignore configured to exclude all its contents from version control. The main contents of the directory are:

    • uncommitted/config_variables.yml, which holds sensitive information, such as database credentials and other secrets.

    • uncommitted/documentation, which contains Data Docs generated from Expectations, Validation Results, and other metadata.

    • uncommitted/validations, which holds Validation Results generated by Great Expectations.

Back in your terminal, go ahead and hit Enter to proceed.