Speakers are not defined.
Check the Unsupervised machine learning with Python website for more details.