Update examples to be v.0.3.0+ compliant #69

ljvmiranda921 · 2017-10-21T09:27:33Z

Hi @reiinakano , here's my PR! 👍

This commit solves Issue #62 and updates Jupyter notebook examples to be compliant with the latest version. Some things to note:

Summary

Demos that used the plotters API were removed. Some examples demonstrate by showing how to do it with the factory, and by the plotters API. Since both of them are deprecated, I just removed the other and focused on the updated way to do it.
Add a .gitignore file to untrack .ipynb_checkpoints I think they're not needed so I have git ignore them.
Explicitly pass a parameter to cmap to avoid matplotlib deprecation warning. This seems a bit weird for me. My environment has matplotlib 2.0.2 and of course scikit-plot 0.3.1. When I call plot_roc_curve or plot_precision_curve, I encounter a deprecation warning on the use of spectral or spectral_r.

So whenever I do this:

import matplotlib.pyplot as plt
from sklearn.naive_bayes import GaussianNB
from sklearn.datasets import load_digits as load_data
import scikitplot as skplt

X, y = load_data(return_X_y=True)
nb = GaussianNB()
nb.fit(X,y)
y_probas = nb.predict_proba(X)

skplt.metrics.plot_precision_recall_curve(y, y_probas)
plt.show()

This appears:

 MatplotlibDeprecationWarning: The spectral and spectral_r colormap was deprecated in version 2.0. Use nipy_spectral and nipy_spectral_r instead.
  warnings.warn(message, mplDeprecation, stacklevel=1)

Although I checked the code in the repo and it's actually supplying nipy_spectral. Hmmm, not sure if it's just an environment problem for me. Either way, I solve it by explicitly passing the cmap parameter instead.

I committed a lot (almost per file), I can rebase my commits after your suggestion/revisions/review.
Thank you so much! Sorry if this PR took me a long time, thesis happened. 😞

This commit adds a .gitignore file to ignore tracking changes in the .ipynb_checkpoints directory.

This commit updates this example to be compliant on v.0.3.0. Using the classifier_factory has been removed, and instead plots the confusion matrix using scikitplot.metrics.plot_confusion_matrix. In addition, the whole example that uses the plotters module has also been removed due to deprecation.

This commit updates the plot_elbow curve example where it removes the need to create a clusterer factory and instead uses the scikitplot.cluster.plot_elbow_curve method.

This commit updates the use of the plot_feature_importance method to be compliant with v.0.3.0. The additional example that uses plotters has been removed for it is deprecated already.

This commit updates the example notebook by using the scikitplot.metrics.plot_ks_statistic API. This also removes the plotter example due to depreciation.

This commit defines the colormap explicitly in the plot_precision_recall example. It turns out that for other matplotlib versions, a warning on nipy_spectral/nipy_spectral_r is being raised. Defining it avoids this ugly warning.

This commit updates plot_roc_curve to be compliant with version 0.3.0. Moreover, the colormap here is defined explicitly in the function call in order to avoid the ugly warning raised by nipy_spectral.

reiinakano · 2017-10-21T13:13:56Z

Hi @ljvmiranda921! Thanks a lot for this great PR! Nothing's wrong with your environment, there's just no released version with the fix for the deprecated colormap yet.

No need to rebase, Github has an option to squash commits upon merge. So far LGTM!

ljvmiranda921 added 14 commits October 21, 2017 12:18

Add .gitignore file

28474f6

This commit adds a .gitignore file to ignore tracking changes in the .ipynb_checkpoints directory.

Update plot_elbow_curve example notebook (#62)

3de8b84

This commit updates the plot_elbow curve example where it removes the need to create a clusterer factory and instead uses the scikitplot.cluster.plot_elbow_curve method.

Update plot_feature_importance notebook (#62)

d0fcde3

This commit updates the use of the plot_feature_importance method to be compliant with v.0.3.0. The additional example that uses plotters has been removed for it is deprecated already.

Update plot_feature_importance

23f2c06

Update plot_ks_statistic notebook example (#62)

b14b0bd

This commit updates the example notebook by using the scikitplot.metrics.plot_ks_statistic API. This also removes the plotter example due to depreciation.

Update plot_learning_curve notebook example (#62)

20897b8

Update plot_pca_2d_projection example notebook (#62)

b51d529

Update plot_pca_2d_projection example notebook (#62)

6a1f26f

Update plot_precision_recall_curve example (#62)

513a9fd

Define colormaps explicitly in examples

a5e32dd

This commit defines the colormap explicitly in the plot_precision_recall example. It turns out that for other matplotlib versions, a warning on nipy_spectral/nipy_spectral_r is being raised. Defining it avoids this ugly warning.

Update plot_roc_curve examples (#62)

da2c39c

This commit updates plot_roc_curve to be compliant with version 0.3.0. Moreover, the colormap here is defined explicitly in the function call in order to avoid the ugly warning raised by nipy_spectral.

Update description on example

c9e8888

Update plot_silhouette notebook example (#62)

9705282

reiinakano merged commit 7d39a7b into reiinakano:master Oct 21, 2017

ljvmiranda921 mentioned this pull request Oct 21, 2017

Merge pull request #1 from reiinakano/master #70

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update examples to be v.0.3.0+ compliant #69

Update examples to be v.0.3.0+ compliant #69

ljvmiranda921 commented Oct 21, 2017

reiinakano commented Oct 21, 2017

Update examples to be v.0.3.0+ compliant #69

Update examples to be v.0.3.0+ compliant #69

Conversation

ljvmiranda921 commented Oct 21, 2017

Summary

reiinakano commented Oct 21, 2017