Tamar Grey

@jacktang Thanks for explaining. I think having a `df_date` datataframe makes a lot of sense. I think what you're describing makes sense to be able to leverage Featuretools with multiple...

One thing of note is that you can apply primitives in different orders and get different sets of primitives, and I think that can be really apparent with transform stacking....

It'd also be useful if this section explained how to define `ColumnSchema` objects for the `return_types` parameter in dfs. We'd want to avoid users specifying column schemas that are too...

@eccabay I want to make sure we're separating out the actual handling of null values from the concept of nullable type incompatibilities. Many of our components can't handle nans but...

When this gets implemented, we should add `BooleanNullable` to the parametrization in `test_arima_regressor_with_nullable_types`

Opened https://github.com/alteryx/woodwork/issues/1670 which will be necessary to properly display the differences in woodwork typing info to users.

@lhoch5813 That import is not coming from a featuretools file. Rather, our call in our `num_peaks.py` file to `from scipy.signal import find_peaks` eventually reaches the import you mention, which is...

Seems like everything works as of catboost 1.1.1 (with support coming over the course of a few prior releases): ```python import pandas as pd from catboost import CatBoostClassifier for dtype...

I think this only happens if y's min value is greater than 0 - it's the call to `y_ww.apply(np.log)` that must maintain the nullable type, singe the apply above it...

We should also think about this with the `TargetImputer`, which would have this same problem