Comment by porker

> No, polars or spark is not a good answer, those are optimized for data engineering performance, not a holistic approach to data science.

Can you expand on why Polars isn't optimised for a holistic approach to data science?