No Pandas Were Harmed: Elegant and Efficient Data Analytics with Polars

Workshop Description

Data analytics doesn’t have to be slow or cumbersome. With Polars↗, a next-generation DataFrame library built for performance and simplicity, you can process massive datasets faster than ever. In this hands-on workshop, you will discover how to leverage Polars to support your data workflows from the fundamentals to the more advanced analytics scenarios.

In the first part of the workshop, we will introduce Polars as the modern library for data analytics in Python, covering its data model, engine design, and how it compares to other analytics tools like Pandas↗, which is effective for smaller datasets, and Apache Spark↗, which is horizontally scalable but inefficient for smaller datasets because of the overhead of the distributed infrastructure.

The second part will dive into a real-world dataset of NYC taxi trips, containing hundreds of millions of rows. We will walk through efficient data ingestion and cleaning, and then explore how to perform lightning-fast queries and groupings directly in a notebook environment. You will also learn how to visualize your findings with simple yet powerful plotting techniques.

Finally, we will tackle more complex and realistic analytics scenarios, such as identifying the top-earning drivers over a specific time window, and visualizing their most profitable trips and neighborhoods. Through these scenarios, we will showcase how Polars empowers you to move seamlessly from raw data to meaningful insights.