Preprocessing

Most datasets need to be preprocessed/transformed before they can be passed to the model. This module includes common transformers that are compatible with sklearn Pipeline or ColumnTransformer.

DateTransformer

Bases: TransformerMixin, BaseEstimator

Transform date features by deriving useful date/time attributes:

• date attributes: Year, Month, Week, Day, Dayofweek, Dayofyear, Is_month_end, Is_month_start, Is_quarter_end, Is_quarter_start, Is_year_end, Is_year_start.
• time attributes: Hour, Minute, Second.

__init__(date_feats=None, time=False, drop=True)

Parameters:

Name	Type	Description	Default
date_feats	Iterable	Date features to transform. If None, all features with datetime64 data type will be used.	None
time	bool	Whether to add time-related derived features such as Hour/Minute/...	False
drop	bool	Whether to drop date features used.	True

fit(X, y=None)

Populate date features if not provided at initialization.

Parameters:

Name	Type	Description	Default
X	DataFrame	Dataframe that has the date features to transform.	required
y	array | DataFrame | None	Included for completeness to be compatible with scikit-learn transformers and pipelines but will not be used.	None

Returns:

Name	Type	Description
self	DateTransformer	Fitted date transformer.

transform(X, y=None)

Derive the date/time attributes for all date features.

Parameters:

Name	Type	Description	Default
X	DataFrame	Dataframe that has the date features to transform.	required
y	array | DataFrame | None	Included for completeness to be compatible with scikit-learn transformers and pipelines but will not be used.	None

Returns:

Name	Type	Description
X_tr	DataFrame	Dataframe with derived date/time features and NaN indicators.