Countplotting a binary Series takes an excessive amount of time, and ends up producing a strange plot

As can be seen in the screenshot below, my y_train Series contains two values: 0 and 1. Why, then, does the countplot take almost 4 minutes to run, and when it finally finishes, it produces a plot that has nothing to do with the data?

You should directly pass the Series to x:

sns.countplot(x=y_train)

why did it fail?

Assuming the latest seaborn (0.13.2).

If you pass a Series to data, this is considered a wide format.

data: DataFrame, Series, dict, array, or list of arrays

Dataset for plotting. If x and y are absent, this is interpreted as wide-form. Otherwise it is expected to be long-form.

import seaborn as sns

y_train = pd.Series([0, 1, 1, 0], name='count')

sns.countplot(data=y_train)

Output:

In your case, this creates more than 20,000 bars.

A workaround could be to convert to_frame and pass the name as x:

import seaborn as sns

y_train = pd.Series(np.repeat([0, 1], [16548, 5251]), name='count')

sns.countplot(data=y_train.to_frame(), x='count')