Using the first 35,000 words is a bit unfair for a rapper such as Lil Wayne who's been releasing work since he was 14.

Also I wonder if this is including proper nouns and other references. (I'd think it should, but it's hard to account for the fact that referencing seven different Chris's would be counted as one token used seven times. Similarly, many words have many meanings, and those are all being lumped together as well, so no accounting here can probably ever be perfect).

If you had all the lyrics for all the rappers I think I'd - aggregate word counts - combine variations - remove most commonly used words in each language (I, I'm, You, You're, etc)

Then see who came out ahead. You shouldn't get penalized for releasing more.

You could probably do a bunch of cool analysis with that data.

edit: Oh no, there's actually a Genius API isn't there. No no no no. I have no time!

▲ Rendello a day ago | parent | next [-]

The original author was pretty clear about the limitations of his work. I certainly would like to see an updated version, so I'm glad you got nerd sniped and not me. I look forward to see your super accurate updated version in a few months ;)

▲ Jach a day ago | parent [-]

I've wanted to see a version that segmented rappers by topic clusters of what they rap about, with fine enough details that you could take something like "drugs" and drill down or aggregate specific ones, and even have some sentiment data as well like pot good/X bad. It'd be fun to see who has the most unique general topics and topics only covered by one rapper. I can see how that might be biased in favor of total lifetime output, but perhaps not. My favorite dead rapper is Eyedea, his album output didn't exceed the 35k word minimum for the vocab list but his topic breadth was pretty wide. I've thought LLMs might be good enough to do that for a lot of songs now, if they don't have a panic attack over the language anyway, but I haven't experimented. Maybe someone else can be sniped into doing it~

▲ defrost a day ago | parent [-]

I too have always been curious about the volumetric comparison of rap lyrics about Edgar Allan Poe Vs those about the people in the front row.

* https://www.youtube.com/watch?v=j8Z0VynTR84

* https://www.youtube.com/watch?v=lqCyTM1bF6Q

▲ Rendello a day ago | parent [-]

Aesop Rock's first on that graph, I think his track about going to the drive-through is pretty poetic:

https://genius.com/Aesop-rock-time-moves-differently-here-ly...

▲ defrost a day ago | parent [-]

Moving forward to the B's there are lyrics that are poetic, specific, expand the vocabulary breadth of rap and don't see much in the way of overlap:

  Brother boys, Yolŋu boys, all the way from Arnhem Land
  räwakpuy yindi djäl napurr dhuwal giritjinyaraw
  Yolŋu balanda buŋgul go
  dhumurr'yurra ŋanya marrtji go
  napurr ga djälthirr
  napurr dhu wiripukum walalaŋ maŋutji
  marr ga manapanmirr wiripu wiripu miṯtji go
  babu'yurrnha ŋanya marrtji go

▲ giancarlostoro a day ago | parent | prev [-]

Lil Wayne also has an insane amount of mixtapes and freestyles.