It would be great to offer Polars support, it is currently half as popular as Pandas, and generally work better for large datasets. Polars is bound to replace most data-scientist day to day operations within the next five years.

Thanks for developing bigframes, it very useful.