Standards Drive
Freedom.

ALL IN ON IBIS

The Standard Dataframe API
for Python

The Ibis project is the new front door for data analytics and machine learning. It fills a critical gap in data analytics. It’s all the plumbing required to connect to different data sources and maintain code.

EMBRACING A FAMILIAR INTERFACE

Enabling Production-Grade Pythonic Analytics with the Flexibility of pandas at Scale

MAINTENANCE WITH EASE

Standardizing Workflows for Manipulating Data

SCALABLE WHERE YOU NEED IT

Go from Code to Dev to Production with Ease

Make Your Codebase Portable

Try Ibis

Your Ibis Code

import ibis
conn = ibis.connect("mydb://...")
t = conn.table('mytable')
(t.filter(t.x > 99)
.group_by('g')
.aggregate(z=t.y.sum())
.sort_by('z')
.execute())
With Ibis graphic full

Databases and Query Engines

Ibis makes code
portability
seamless.

One line of code connects to countless backends.



import ibis
ibis.<backend>.connect(config)
lines bottom
Start Logo
Apache Arrow Data Fusion ↗
Apache Impala ↗
ClickHouse ↗
Dask ↗
DuckDB ↗
Google Big Query ↗
MS SQL Server ↗
MySQL ↗
Pandas ↗
Polars ↗
PostgreSQL ↗
PySpark ↗
Snowflake ↗
SQLite ↗
Trino ↗
Back Logo

Ibis adoption is
taking off

GitHub Starts

GitHub Stars

4.0k

Commits

Commits

+7,800

Contributors / Maintainers

Contributors / Maintainers

158

Active Projects

Active Projects

1,100

Ibis in
leading enterprises


Two Sigma Google Microsoft

Get into Ibis

Ibis

RESOURCES

Read

Blogs about Ibis

When to use Ibis over pandas

Ibis tackles massive datasets more efficiently than pandas. As a consistent Python API, it pares down large datasets before pulling what you need into memory. Learn more about how Ibis stacks up against pandas.

+
Achieve code portability and performance with Ibis

No matter the engine you’re using to wrangle data, the Ibis syntax and concepts stay the same. Gain insight into how Ibis can increase code portability, and the performance gains you’ll get, with this 101.

+
Engine agnostic analytics with Ibis

Ibis can solve technical debt and organizational issues that arise from maintaining a python-centric, SQL-heavy, memory-intensive analytics code base. Learn how its lightweight interface can be a boon for analytics teams.

+
Through one Python API, Ibis connects to 15+ backends

Learn how Ibis can give a single developer power over data movement and migration - all without having to understand how to manipulate the programming infrastructures of the data sources they need to connect to.

+
Analyze a massive dataset with BigQuery and Ibis

Go in-depth: We take a Hacker News dataset from Google BigQuery and use Ibis to analyze the platform’s most common topics – without needing to pull the dataset’s 33 million rows into local memory.

+
Use Ibis + Substrait to supercharge code portability

Using the Ibis interface and Substrait intermediate representation for language, we demonstrate how Substrait supercharges portability by testing on the Apache Arrow DataFusion backend. This leads to ease of use – and innovation.

+

Watch

Ibis talks from The Data Thread

What is Ibis simple demo

TESTIMONIALS

What people are saying

quote icon

@DuckDB 🤝 @IbisData! Like using a DataFrame API, but want 8x the speed of Pandas? Use Ibis with the new DuckDB backend! Now you can use the Ibis API both for querying remote DBs and super fast local analysis. You can always mix and match with SQL too!


AlexMonahan @__AlexMonahan__ Twitter
quote icon

VAST's Sigma frontend is the best example for why we need something like@IbisData 😊 standardized query plan representation) for #infosec. #dataengineering leads the way. ⚙️


@mavam @mavam Twitter
quote icon

Exciting to see the accumulated UX improvements in authoring complex analyses in @IbisData — significant productivity improvements over both SQL and pandas in many places. 👉 Here @cpcloudy analyzes the project’s CI performance data in BigQuery.


@wesmckinn @wesmckinn Twitter

Voltron Data offers support for developers and enterprises working with Ibis.

We understand that working with open source tools can sometimes be difficult and challenging. From answering technical issues on GitHub to partnering on developing core system technology, we are here to help as little and as much as you require. Please reach out to hello@voltrondata.com with any questions or needs you may have.