There is no real connection between the two here, this data analysis is just fluff to me. The sample size of the two datasets are differ to a great amount, also the content and structure of lyrics versus a news articles are in no way similar and RG is trying to force a connection that is not there.