Plot Pipeline Display¶
================================================================= Displaying Pipelines¶
The default configuration for displaying a pipeline in a Jupyter Notebook is
'diagram' where set_config(display='diagram'). To deactivate HTML representation,
use set_config(display='text').
To see more detailed steps in the visualization of the pipeline, click on the steps in the pipeline.
Imports for Interactive Pipeline Diagram Rendering¶
Scikit-learn’s HTML diagram display mode renders pipelines and composite estimators as interactive, expandable diagrams directly in Jupyter notebooks: Controlled via set_config(display='diagram') (the default) or set_config(display='text'), the diagram mode generates an HTML/CSS representation where each pipeline step appears as a clickable box – expanding it reveals the estimator’s parameters. This is particularly useful for complex nested structures like a ColumnTransformer inside a Pipeline inside a GridSearchCV, where the text representation would be a wall of nested parentheses but the diagram clearly shows the data flow from preprocessing branches through to the final estimator.
The examples progress from simple linear pipelines to deeply nested composite estimators, demonstrating how the visualization scales with complexity: A basic StandardScaler -> LogisticRegression pipeline renders as two connected boxes; a ColumnTransformer with separate categorical and numerical preprocessing branches renders as a parallel structure with converging arrows; and a GridSearchCV wrapping the entire pipeline shows the search object containing the pipeline diagram. The make_pipeline convenience function infers step names from class names, while the explicit Pipeline(steps=[...]) constructor allows custom naming. These diagrams serve as both documentation and debugging tools – verifying at a glance that features are routed to the correct transformers and that the estimator hierarchy matches the intended design.
# Authors: The scikit-learn developers
# SPDX-License-Identifier: BSD-3-Clause
# %%
# Displaying a Pipeline with a Preprocessing Step and Classifier
# ##############################################################
# This section constructs a :class:`~sklearn.pipeline.Pipeline` with a preprocessing
# step, :class:`~sklearn.preprocessing.StandardScaler`, and classifier,
# :class:`~sklearn.linear_model.LogisticRegression`, and displays its visual
# representation.
from sklearn import set_config
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler
steps = [
("preprocessing", StandardScaler()),
("classifier", LogisticRegression()),
]
pipe = Pipeline(steps)
# %%
# To visualize the diagram, the default is `display='diagram'`.
set_config(display="diagram")
pipe # click on the diagram below to see the details of each step
# %%
# To view the text pipeline, change to `display='text'`.
set_config(display="text")
pipe
# %%
# Put back the default display
set_config(display="diagram")
# %%
# Displaying a Pipeline Chaining Multiple Preprocessing Steps & Classifier
# ########################################################################
# This section constructs a :class:`~sklearn.pipeline.Pipeline` with multiple
# preprocessing steps, :class:`~sklearn.preprocessing.PolynomialFeatures` and
# :class:`~sklearn.preprocessing.StandardScaler`, and a classifier step,
# :class:`~sklearn.linear_model.LogisticRegression`, and displays its visual
# representation.
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import PolynomialFeatures, StandardScaler
steps = [
("standard_scaler", StandardScaler()),
("polynomial", PolynomialFeatures(degree=3)),
("classifier", LogisticRegression(C=2.0)),
]
pipe = Pipeline(steps)
pipe # click on the diagram below to see the details of each step
# %%
# Displaying a Pipeline and Dimensionality Reduction and Classifier
# #################################################################
# This section constructs a :class:`~sklearn.pipeline.Pipeline` with a
# dimensionality reduction step, :class:`~sklearn.decomposition.PCA`,
# a classifier, :class:`~sklearn.svm.SVC`, and displays its visual
# representation.
from sklearn.decomposition import PCA
from sklearn.pipeline import Pipeline
from sklearn.svm import SVC
steps = [("reduce_dim", PCA(n_components=4)), ("classifier", SVC(kernel="linear"))]
pipe = Pipeline(steps)
pipe # click on the diagram below to see the details of each step
# %%
# Displaying a Complex Pipeline Chaining a Column Transformer
# ###########################################################
# This section constructs a complex :class:`~sklearn.pipeline.Pipeline` with a
# :class:`~sklearn.compose.ColumnTransformer` and a classifier,
# :class:`~sklearn.linear_model.LogisticRegression`, and displays its visual
# representation.
import numpy as np
from sklearn.compose import ColumnTransformer
from sklearn.impute import SimpleImputer
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import Pipeline, make_pipeline
from sklearn.preprocessing import OneHotEncoder, StandardScaler
numeric_preprocessor = Pipeline(
steps=[
("imputation_mean", SimpleImputer(missing_values=np.nan, strategy="mean")),
("scaler", StandardScaler()),
]
)
categorical_preprocessor = Pipeline(
steps=[
(
"imputation_constant",
SimpleImputer(fill_value="missing", strategy="constant"),
),
("onehot", OneHotEncoder(handle_unknown="ignore")),
]
)
preprocessor = ColumnTransformer(
[
("categorical", categorical_preprocessor, ["state", "gender"]),
("numerical", numeric_preprocessor, ["age", "weight"]),
]
)
pipe = make_pipeline(preprocessor, LogisticRegression(max_iter=500))
pipe # click on the diagram below to see the details of each step
# %%
# Displaying a Grid Search over a Pipeline with a Classifier
# ##########################################################
# This section constructs a :class:`~sklearn.model_selection.GridSearchCV`
# over a :class:`~sklearn.pipeline.Pipeline` with
# :class:`~sklearn.ensemble.RandomForestClassifier` and displays its visual
# representation.
import numpy as np
from sklearn.compose import ColumnTransformer
from sklearn.ensemble import RandomForestClassifier
from sklearn.impute import SimpleImputer
from sklearn.model_selection import GridSearchCV
from sklearn.pipeline import Pipeline, make_pipeline
from sklearn.preprocessing import OneHotEncoder, StandardScaler
numeric_preprocessor = Pipeline(
steps=[
("imputation_mean", SimpleImputer(missing_values=np.nan, strategy="mean")),
("scaler", StandardScaler()),
]
)
categorical_preprocessor = Pipeline(
steps=[
(
"imputation_constant",
SimpleImputer(fill_value="missing", strategy="constant"),
),
("onehot", OneHotEncoder(handle_unknown="ignore")),
]
)
preprocessor = ColumnTransformer(
[
("categorical", categorical_preprocessor, ["state", "gender"]),
("numerical", numeric_preprocessor, ["age", "weight"]),
]
)
pipe = Pipeline(
steps=[("preprocessor", preprocessor), ("classifier", RandomForestClassifier())]
)
param_grid = {
"classifier__n_estimators": [200, 500],
"classifier__max_features": ["auto", "sqrt", "log2"],
"classifier__max_depth": [4, 5, 6, 7, 8],
"classifier__criterion": ["gini", "entropy"],
}
grid_search = GridSearchCV(pipe, param_grid=param_grid, n_jobs=1)
grid_search # click on the diagram below to see the details of each step