## How can I change the size of a Yellowbrick plot?¶

You can change the size of a plot by passing in the desired dimensions in pixels on instantiation of the visualizer:

# Import the visualizer

# Instantiate the visualizer using the size param
classes=classes, features=features, size=(1080, 720)
)

...


Note: we are considering adding support for passing in size in inches in a future Yellowbrick release. For a convenient inch-to-pixel converter, check out www.unitconversion.org.

## How can I change the title of a Yellowbrick plot?¶

You can change the title of a plot by passing in the desired title as a string on instantiation:

from yellowbrick.classifier import ROCAUC
from sklearn.linear_model import RidgeClassifier

my_title = "ROCAUC Curves for Multiclass RidgeClassifier"

# Instantiate the visualizer passing the custom title
visualizer = ROCAUC(
RidgeClassifier(), classes=classes, title=my_title
)

...


## How can I change the color of a Yellowbrick plot?¶

To customize coloring in your plot, use the colors or cmap (or colormap) arguments. Note that different visualizers may require slightly different arguments depending on how they construct the plots.

For instance, the Manifold Visualization accepts a colors argument, for which discrete targets should be the name of one of the Colors and Style or a list of matplotlib colors represented as strings: For instance, the Manifold Visualization accepts a colors argument, for which discrete targets should be the name of a palette from the Yellowbrick Colors and Style or a list of matplotlib colors represented as strings:

from yellowbrick.features.manifold import Manifold

visualizer = Manifold(
manifold="tsne", target="discrete", colors=["teal", "orchid"]
)

...


… whereas for continuous targets, colors should be a colormap:

from yellowbrick.features.manifold import Manifold

visualizer = Manifold(
manifold="isomap", target="continuous", colors="YlOrRd"
)

...


Other visualizers accept a cmap argument:

from sklearn.linear_model import LogisticRegression
from yellowbrick.classifier import ConfusionMatrix

visualizer = ConfusionMatrix(
LogisticRegression(), cmap="YlGnBu"
)

...


Or a colormap argument:

from yellowbrick.features import ParallelCoordinates

# Instantiate the visualizer
visualizer = ParallelCoordinates(
classes=classes, features=features, colormap="PRGn"
)

...


The Residuals Plot accepts color argument for the training and test points, train_color and test_color, respectively:

from yellowbrick.regressor import ResidualsPlot
from sklearn.linear_model import ElasticNet

visualizer = ResidualsPlot(
model=ElasticNet()
train_color=train_color,  # color of points model was trained on
test_color=train_color,   # color of points model was tested on
line_color=line_color    # color of zero-error line
)


## How can I save a Yellowbrick plot?¶

Save your Yellowbrick plot by passing an outpath into poof():

from sklearn.cluster import MiniBatchKMeans
from yellowbrick.cluster import KElbowVisualizer

visualizer = KElbowVisualizer(MiniBatchKMeans(), k=(4,12))

visualizer.fit(X)
visualizer.poof(outpath="kelbow_minibatchkmeans.png")

...


Most backends support png, pdf, ps, eps and svg to save your work!

## How can I make overlapping points show up better?¶

You can use the alpha param to change the opacity of plotted points (where alpha=1 is complete opacity, and alpha=0 is complete transparency):

from yellowbrick.contrib.scatter import ScatterVisualizer

visualizer = ScatterVisualizer(
x="light", y="C02", classes=classes, alpha=0.5
)


## How can I access the sample datasets used in the examples?¶

Visit the Example Datasets page.