Always know what to expect from your data! #expectgreatdata
What is Great Expectations?¶
Great Expectations is a Python-based open-source library for validating, documenting, and profiling your data. It helps you to maintain data quality and improve communication about data between teams.
Software developers have long known that automated testing is essential for managing complex codebases. Great Expectations brings the same discipline, confidence, and acceleration to data science and data engineering teams.
Why would I use Great Expectations?¶
One of the key statements we hear from data engineering teams that use Great Expectations is: “Our stakeholders would notice data issues before we did – which eroded trust in our data!”
With Great Expectations, you can assert what you expect from the data you load and transform, and catch data issues quickly – Expectations are basically unit tests for your data. Not only that, but Great Expectations also creates data documentation and data quality reports from those Expectations. Data science and data engineering teams use Great Expectations to:
Test data they ingest from other teams or vendors and ensure its validity.
Validate data they transform as a step in their data pipeline in order to ensure the correctness of transformations.
Prevent data quality issues from slipping into data products.
Streamline knowledge capture from subject-matter experts and make implicit knowledge explicit.
Develop rich, shared documentation of their data.
You can read more about how data teams use Great Expectations in our case studies.
Expectations are assertions about your data. In Great Expectations, those assertions are expressed in a declarative language in the form of simple, human-readable Python methods. For example, in order to assert that you want the column “passenger_count” to be between 1 and 6, you can say:
expect_column_values_to_be_between(column="passenger_count", min_value=1, max_value=6)
Great Expectations then uses this statement to validate whether the column
passenger_countin a given table is indeed between 1 and 6, and returns a success or failure result. The library currently provides several dozen highly expressive built-in Expectations, and allows you to write custom Expectations.
Automated data profiling
Writing pipeline tests from scratch can be tedious and overwhelming. Great Expectations jump starts the process by providing automated data profiling. The library profiles your data to get basic statistics, and automatically generates a suite of Expectations based on what is observed in the data.
For example, using the profiler on a column
passenger_countthat only contains integer values between 1 and 6, Great Expectations automatically generates this Expectation we’ve already seen:
expect_column_values_to_be_between(column="passenger_count", min_value=1, max_value=6).
This allows you to quickly create tests for your data, without having to write them from scratch.
Once you’ve created your Expectations, Great Expectations can load any batch or several batches of data to validate with your suite of Expectations. Great Expectations tells you whether each Expectation in an Expectation Suite passes or fails, and returns any unexpected values that failed a test, which can significantly speed up debugging data issues!
Great Expectations renders Expectations to clean, human-readable documentation, which we call Data Docs, see the screenshot below. These HTML docs contain both your Expectation Suites as well as your data validation results each time validation is run – think of it as a continuously updated data quality report.
Support for various Datasources and Store backends
Great Expectations currently supports native execution of Expectations against various Datasources, such as Pandas dataframes, Spark dataframes, and SQL databases via SQLAlchemy. This means you’re not tied to having your data in a database in order to validate it: You can also run Great Expectations against CSV files or any piece of data you can load into a dataframe.
Great Expectations is highly configurable. It allows you to store all relevant metadata, such as the Expectations and validation results in file systems, database backends, as well as cloud storage such as S3 and Google Cloud Storage, by configuring metadata Stores.
What does Great Expectations NOT do?¶
Great Expectations is NOT a pipeline execution framework.
Great Expectations is NOT a data versioning tool.
Great Expectations currently works best in a Python environment.
Great Expectations is Python-based. You can invoke it from the command line without using a Python programming environment, but if you’re working in another ecosystem, other tools might be a better choice. If you’re running in a pure R environment, you might consider assertR as an alternative. Within the TensorFlow ecosystem, TFDV fulfills a similar function as Great Expectations.
How do I get started?¶
Check out Getting started with Great Expectations to set up your first local deployment of Great Expectations, and learn important concepts along the way.
If you’d like to contribute to Great Expectations, please start here.
If you’re interested in a paid support contract or consulting services for Great Expectations, please see options here
For other questions and resources, please visit Community resources.