I’ve had this project finished for more than a month but I just can’t bring myself to do a writeup about it…
Is this like a recipe blog where you’d rather I just get to the pretty data viz without prattling on about where the idea came from and what it took to make it happen?
Please let me know, I’d appreciate it!
In the meantime, though, allow me to indulge in a little prattling.
This project was made by request of my boyfriend. He’d been noticing a lot of Star Wars references in rap, and wondered what the most mentioned character was.
This idea sounded super fun–and someone else already had it. It turned out that Wired had answered the question in 2015.
Of course, we could have proceeded to see if any new Star-Wars-referencing music was released in the intervening five years, but seeing that the idea already had been done took the wind out of my sails.
My boyfriend proposed we do something else instead: which rapper mentions the most food?
Even better, the idea came with a catchy title built in: Who Is the Hungriest Rapper?
To find the hungriest rapper, I needed a list of foods to search for.
This is the crux of the entire project. Any issues here will percolate down through the rest of the analysis. However, listing out every single food a rapper might mention is extraordinarily difficult, especially for just one person.
I came up with a few ground rules to keep myself sane.
- All I can do is my best. My list of foods will certainly be incomplete, but if I put in a good-faith effort to gather every food I can find, that’ll have to be good enough.
- I relied on Wikipedia heavily, as it documents tons of types of food across many cultures.
- I also manually read over sample lyrics to find things I had missed on my first go-round.
- Slang terms count as food mentions if they have a culinary origin
- For example, beef, as in an argument or dispute, counts as a food mention. It may come from disputes between cattle farmers over who got the slaughtered remains of cows.
- It is OK to exclude certain words if they almost always refer to a non-food thing and appear in more than a few thousand lyrics.
- For this reason, I excluded words like roll and jam as they almost never referred to the food and appeared too frequently to audit with losing my mind
- Foods used in compound words count individually
- e.g. “bean burrito” counts as both “bean” and “burrito”
- Metaphorical references to food don’t count if they don’t use a specific food word
- I can’t possibly catch all these, so I didn’t attempt it
In all, I gathered 822 food words (plus their alternate spellings and plurals). Those led me to more than 270,000 lyrics, of which about 20,000 had to manually reviewed.
My final dataset contained over 180,000 lyrics by more than 7,000 artists.
The hungriest rapper
So who wins the title of Hungriest Rapper?
The answer: it depends! There’s a few ways to decide who mentions the most food, so let’s take a look at three of them.
Average number of foods mentioned per song
This is my favorite method! In my opinion, a truly hungry rapper is one who mentions lots of different kinds of food in all of his or her songs.
Anyone who’s listened to Hail Mary Mallon won’t be surprised by their top ranking on this list–not only does the group contain the famously wordy Aesop Rock, they’re named after the infamous typhoid-spreading cook Mary Mallon.
Percent of songs that mention food
Another method is to simply look at how many of the artist’s songs mention food at all.
Hail Mary Mallon still ranks quite high by this method. Many of the other artists, I believe, are on this list because they mention their own food-related names in their songs.
Still, naming yourself after a food is still a solid sign of hungriness!
Total number of foods mentioned
This, perhaps, is the simplest way to determine the hungriest rapper, but it’s biased towards rappers with longer careers. Regardless, Lil Wayne’s lead over the rest of the pack is quite remarkable! He doesn’t tend to mention a lot of foods per song, but he’s covered quite a diverse array of comestibles all the same.
Most mentioned foods
Since I went to all this trouble to find the hungriest rapper, it seems a shame not to dig a little further into the data.
Most pressingly, what’s the most rapped-about type of food?
A rapper’s food pyramid
The chart above looks like unhealthy foods are more likely to be mentioned. What does a rapper’s diet (as judged by the foods they mention) look like compared to a healthy diet?
Move the slider to compare a rapper’s food pyramid with the UK Government’s Eat Well food pyramid.
How did I make this?
I thought about writing up a long article, complete with code, to describe my data pipeline.
That felt like it would take too long, so instead I spent far longer making a handy little GIF that oversimplifies the process.
Awesome visualizations as always! 🙂 Is the code being posted anywhere .. particularly interested in that animation! 😀
Thank you! I actually made the animation manually in Photoshop. Is there a particular piece of code you want to see, or just all of it? (Admittedly it’s a little messy!)
As is cool … code + data! Don’t want to waste your time.
That animation is awesome … gosh, can you share the photoshop file! 😀 … never knew you could make that kind of stuff. How long did it take you!?
Here you go, I cleaned it upfrom 8 messy R files!
The animation probably took me 3-4 hours over a couple of days. It’s not difficult but took some planning–I made all the moving pieces the same height so I could go into the animation window and move them all together about down 20 pixels per frame
I’m in favor of prattle, it makes the data more interesting.
Next you can look at the Hungaryest rappers! https://en.wikipedia.org/wiki/Hungarian_hip_hop
Amazing data collection.
Never thought that you could collect data on the hungriest rappers too 😂.
I got to know about your blog on yesterday’s MIT Technology Review.
I’m glad I checked it out.
And the animation is great too
+1 for the story behind the analysis!
Thanks for your post 🙂
[…] Who is the hungriest rapper? Erin Davis set out to find the answer using data from Spotify and Genius, counting up who mentions food the most. It’s a fun article, but the bit I liked best comes at the very end. Davis has put together a little animated gif showing her data pipeline as a literal pipeline, with lyrics pouring in at the top and a Spotify funnel filtering out all the rappers. It somehow looks exactly like my brain feels when I’ve figured out all the steps in an analysis project. […]
That is a really fun project! Next time you are tempted to use Wikipedia as a source for machine readable data I’d suggest taking a look at Wikidata which is the machine readable sister project. You could for example use this query: https://w.wiki/soF to get everything that the Wikidata community labelled as a food (or a subclass thereof) and added an English label to it. This gives you a lot more words to search for – probably too many for your case – but I thought it might still be interesting to know as an alternative for future projects 🙂
Thank you, that’s super useful! I’ve dabbled (very lightly) with Wikidata but didn’t even think of it for this project. There’s always next time 🙂