Use "vegafusion" data transformer in Altair 5.1 instead of vf.enable()

[jonmmease]

This course looks great!

We're working on moving VegaFusion's Altair integration into Altair itself, and as of Altair 5.1 the functionality of vf.enable() can be replaced by:

alt.data_transformers.enable("vegafusion")

This avoids the need to import VegaFusion separately, and is what is now recommended in the Altair documentation: https://altair-viz.github.io/user_guide/large_datasets.html#vegafusion-data-transformer. It will also cause Altair to invoke VegaFusion when saving charts with chart.save() (replacing VegaFusion's vf.save_* functions).

The full interactive functionality of vf.enable_widget() isn't integrated into Altair yet, so for now vf.enable_widget() is still the best approach for use cases that involve interactive filtering of large datasets.

Another Altair 5.1 feature that might be useful for teaching is the new chart.transformed_data() method. This uses VegaFusion to make it possible to access the transformed data from a chart as a pandas DataFrame. It doesn't support every transform type yet, but it's already pretty useful for inspecting what most of the common transforms do.