I know Polars is fast, but my data pipelines are written in pandas!
09-08, 14:05–14:40 (Europe/Lisbon), Auditorium

We all know it by now: Polars is blazingly fast™️. Yet my pipelines are all written in pandas, and it will just take too much time to rewrite them in Polars... won't it? Turns out, it takes less than thirty minutes to tame this new arctic beast!


How long do you think it would take to rewrite your data pipelines from pandas to Polars? It turns out, less than you would expect! Of course, "if it ain't broke, don't fix it" - yet there are some fruits that are just hanging too low for you to ignore.

Starting from I/O, to (almost) zero-copy conversion from pandas to Polars, you will quickly realise how cheap and convenient it is to swap some bits of your pipelines from pandas to Polars. Though pandas' API is incredibly good, you will soon realise how Polars took it to the next level and made it much more powerful, expressive and intuitive.

Come for the speed, stay for the syntax!

📍 Keynote outline

  • Polars: the cheapest ways to reap its benefits.
  • Blazingly fast I/O with LazyDataFrames.
  • Powerful column selection with the new selectors module.
  • df.filter(): no more "setting a view vs a copy" warnings.
  • How to write expressive groupby statements and window functions.
  • Nested data? Not a problem!
  • Work with datasets larger than memory.
  • Don't like the syntax? Just use SQL - it even works from the CLI!

👋 Hello there! I am a Machine Learning Engineer at Futura, an edtech startup, and I have been giving lectures in machine learning at the University of Milan for a couple of years. I am also an organiser of Python Milano, Milan's local Python meetup. I love writing sound, well-documented code, time-series forecasting and dataframe libraries.