def scaleDataColumns(df):
"""
Scale columnwise, only certain columns
:param df: dataframe, no scaled data
:return: dataframe, with scaled data
"""
enc = ColumnTransformer(remainder="passthrough",
transformers=[("std", StandardScaler(), [
"ts_listen", "media_id", "release_date",
"media_duration", "user_id", "artist_id"
])])
df_scaled = enc.fit_transform(df)
df = pd.DataFrame(df_scaled, columns=enc.get_feature_names_out())
return df
# :meth:`compose.ColumnTransformer.get_feature_names_out` is avaliable to
# combine feature names of its transformers:
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OneHotEncoder
import pandas as pd
X = pd.DataFrame({"pet": ["dog", "cat", "fish"], "age": [3, 7, 1]})
preprocessor = ColumnTransformer(
[
("numerical", StandardScaler(), ["age"]),
("categorical", OneHotEncoder(), ["pet"]),
],
verbose_feature_names_out=False,
).fit(X)
preprocessor.get_feature_names_out()
# %%
# When this ``preprocessor`` is used with a pipeline, the feature names used
# by the classifier are obtained by slicing and calling
# :term:`get_feature_names_out`:
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import make_pipeline
y = [1, 0, 1]
pipe = make_pipeline(preprocessor, LogisticRegression())
pipe.fit(X, y)
pipe[:-1].get_feature_names_out()
##############################################################################
