Copyright by Gabor Szabo http://szabgab.com, 2015, 2016. All Rights Reserved.

The CMOS podcasts (listed here http://code-maven.com/cmos ) have a transcription.

The format of the transcription looks like this:

<transcript>
  <szabgab host1 Gabor Szabo>
  <foobar guest1 Foo Bar>

  <entry 0:00 szabgab>
    Hello!
  </entry>

  <entry 0:25 foobar>
  </entry>

  ....

</transcript>

Everything is within the <transcription></transcription> tags. At the beginning we declare the participants:

<szabgab host1 Gabor Szabo> declares that the nickname 'szabgab' is used for the host whose real name is 'Gabor Szabo.' <foobar guest1 Foo Bar> declares that the nickname 'foobar' is used for the guest, whose real name is 'Foo Bar.'

Then every time the speaker switches we have an <entry></entry> and within that entry we have the text the person said. When another person starts to speak we create a new <entry></entry> section.

In the opening of each <entry> we write the timestamp of when it starts in the podcast and then the nickname of the person who speaks.

ipython nbconvert --to html ../kaggle-NoShowAppointments/No-show_unbalanced_data.ipynb
mv ../kaggle-NoShowAppointments/No-show_unbalanced_data.html static/html/no-show-unbalanced-data.html

Then create a regular article that links to that page.