Food From The Equator Tastes +28% Better (Unless You’re Rich)

“What do you want for dinner tonight?” I don’t know about you, but that’s pretty much the nightly argument debate at my house. My wife likes (sophisticated, modern) Italian and American, I like (spicy, cheap) Thai and Indian. So, what’s a guy to do? Look at the data, of course!

After analyzing 428,580 reviews by 58,106 people over 10,215 restaurants, we found that food from the equator tastes measurably better. The closer to the equator you get, the higher you see the Yelp ratings tick up. By the time you arrive in Indonesia from Norway, the food tastes 28% better.

food-tastes-better-near-the-equator.pngMore precisely, we found that:

  • Every 2,000 miles you drive south, people rate the regional cuisine 8% higher (0.265 stars) on Yelp.
  • The only cuisine with an average Yelp rating less than 3.0 is Scandinavian. Have you ever had lutefisk? Yeah.

And this isn’t just a one-off or some super-popular cuisine throwing everything off. You can see the same exact pattern happening on every continent. No matter where in the world you go, people think the food tastes better near the equator.

everywhere-you-go-food-tastes-better-near-equator.png

What’s Going On?

Ingredients + Spices = Flavor

What’s going on? Well, no one can say for certain, but here’s my best guess: tropical climates have far greater biodiversity — in the tropical rainforests of Borneo, scientists have documented over 15,000 plant species, including over 2,500 species of orchids! In other words, ingredients. Lots of ingredients.

But, also: bugs. (As anyone who’s been to India during the summer knows, “biodiversity” is just a nice word for the biggest, baddest bugs you’ve ever seen.) So many bugs. Food goes bad way faster in tropical climates, so people had to come up with new and better ways of keeping food edible and tasty. AKA spices — lots of spices.

Intuitively, this makes a lot of sense. But, let’s at least verify that anecdotally. Here are the number of ingredients in some typical and/or traditional fish recipes for a few cuisines:

DishCountry of OriginCountry Latitude# of Ingredients
LutefiskNorway61.9242
Grilled SalmonUnited States37.7757
Tilapia CurryIndia22.57314

Except For Rich Folks

That said, there’s one big exception though: when you restrict to upscale restaurants on Yelp ($$$ and $$$$), there’s basically no real pattern:

it-doesnt-matter-if-youre-rich.png

Why? Well, if you’re rich, location doesn’t really matter. If you’re willing to spend big money on food, you can source good ingredients and sophisticated flavor from anywhere and make good food.

Indeed, the European countries that had the biggest colonial empires had the best upscale food: France, Britain, Spain and Portugal. (In contrast, the European countries that had the best casual food were Turkey and Israel.) It just so happens that Britain and France are actually fairly far north (Paris is actually further north than Minneapolis or Bangor), which explains why European upscale food tastes better in the north.

Of course, if you couldn’t afford the fancy spices and exotic ingredients coming off the ships, you kept eating cold porridge and it didn’t really matter.

Methodology

(Mainly for data nerds. If you aren’t one, skip this section unless you’re looking for a cure for late-night insomnia.)

After downloading the Yelp Dataset, I extracted the restaurant reviews written by users with 3+ reviews across 3+ cuisines and then annotated with cuisine latitude, continent of origin, and restaurant priciness. Next, I normalized each rating by user, continent and priciness to try to isolate user-specific preferences between comparable cuisines. Finally, I calculated an ensemble of weighted OLS regressors against cuisine popularity by continent and priciness and plotted the regressed models against latitude. I did all of the above using Python, pandas, sklearn and bokeh.

One thought on “Food From The Equator Tastes +28% Better (Unless You’re Rich)

  1. Kushal,

    This post is amazing! Who would have thought that taste doesn’t matter to rich people 😛 (joking)

    As an engineer, I love reading articles with data science in it but would never attempt to do this on my own.

    Realistically though, how long does it take you to make these articles if you are compiling all that info? Is my dream hopeless?

Leave a Reply to Eyram Sotome Cancel reply