A Pandas series has over 400 attributes! You can work directly on the column to perform a wide variety of operations.

In Polars, while there is a series object, we generally work with column expressions. These have over 250 attributes.

NumPy and Pandas use 'boolean arrays' for filtering.

Polars uses the .filter method.

Both mechanisms work. An advantage of the Polars approach (that the Pandas .query method replicates) is that it works on the current state of the dataframe in a chain of operations.

What is your favorite Python feature? 🐍🤔

Is there any Visual Studio Code expert out there who can enlighten me on why this is pegging my CPU for 10 seconds whenever I try to run a Project Jupyter cell?

One very convenient feature of Pandas is the ability to read CSV files inside of a ZIP file.

With Polars we need to jump through some hoops.

Sad that so many unmasked Python folks are here, yet Pycon is a masked conference. 🤔🤯

Free 'universities' over time:

📚Libraries
🎥YouTube
🐦Twitter
💪LinkedIn

Have you gotten a degree from any of them?

Stop scrolling and start practicing.

Both Pandas and Polars support 'categorical data'.

Leverage this to reduce memory usage and speed up the computation.

One gripe I have with Pandas is automatic conversion of the index to categories for bar plots.

Here's my code to get around this.

linkedin.com/learning/getti…

Looking for practical data science books?
🎓Ace the Interview by Nick Singh | The Data Science Guy 📕
🚀ML by Sebastian Raschka \n🥇Kaggle by Konrad Banachewicz Luca Massaron
🐼Pandas by Matt Harrison