I was looking at the merge part of the Pandas lecture in the QE for data science series.
The airline example part has a code that does not work: The airline code does not match with the descriptions.
It looks like replacing

avg_delays_w_code = avg_delays.join(carrier_code )

 by

avg_delays_w_code = avg_delays.merge(carrier_code, left_index = True, right_on = 'Code')

does the trick.