In the following notebook, there is a possible mistake in a code cell:

07-how-random-forests-really-work.ipynb

The code cell in question:

pd.DataFrame(dict(cols=trn_xs.columns, imp=m.feature_importances_)).plot('cols', 'imp', 'barh');

The description around the cell denotes that this should plot the feature_importances_ of the RandomForestClassifier, but m is the DecisionTreeClassifier initialized in one of the code cells above, like so:

m = DecisionTreeClassifier(max_leaf_nodes=4).fit(trn_xs, trn_y); 

It is my understanding that the code cell needs to be the following:

pd.DataFrame(dict(cols=trn_xs.columns, imp=rf.feature_importances_)).plot('cols', 'imp', 'barh');

Context for rf:

rf = RandomForestClassifier(100, min_samples_leaf=5)

I would be happy to open a PR, once someone acknowledges that this is in fact a valid correction. Thanks! 🤗

Below is a screenshot to show that they indeed plot different things