tl;dr: Some old but very pretty and satisfying visualizations for a pointless analysis. (JS, JQuery, HighCharts)

At some point I really enjoyed Muse, and then less so. Earlier, before semantic-everything, I wanted to see if I could correlate it with a hypothesis: the lyrics were getting more vague about “you” and “me” and “us”. Plus I wanted an excuse to learn some basic data processing/visualization for web.

Using the Genius API, got the lyrics. Processed them with these filters:

firstPString = '\\b(I|me|my|mine|myself|we|us|our|ours|ourselves)\\b'
secondPString = '\\b(you|your|yours|yourself|yourselves)\\b'
thirdPString = '\\b(they|them|their|themselves)\\b'

And plot using highcharts (mouseover/click around!).

Generally it looks like the variance in word count and pronount count increases with newer albums. How about a different visualization, with the pronoun count ratio?

Hmm, looks like the lyrics are wordier, but I can’t generally see a correlation with songs I enjoy and if they’re more concise or less concrete.

If I Come Back to This

Maybe I’ll update this with my ratings of each song and see if any of these features correlate! For now I made some pretty plots.

Happy to share the Python scripts if you want to do something related, feel free to reach out. I should put them somewhere, but it’s small enough that I don’t think it warrants it’s own repo, but maybe fleshing out this blog post with code snippets? Hmm…