What's in two million recipes

We read 2.2 million recipes from American recipe websites to find the dishes people write down most, the handful of ingredients nearly all of them share, and what you could cook tonight with what's already in your kitchen.

Jump to the recipe finder A Data Drop by sheets.works

Every line behind this box is the name of a real recipe, picked at random from the collection, names in all.

The collection is RecipeNLG, 2,231,142 recipes that researchers at Poznań University of Technology gathered from sites like Food.com, Allrecipes, Epicurious and an old site called cookbooks.com, whose recipes read as if they were typed up from family and church cookbooks. After removing copies of the same recipe there are 2,089,091 left.

Casseroles

The names in red have the word casserole in them. Across the whole collection, 80,564 recipe names do, which is about one recipe in every 26.

Chicken casserole

Every name on screen now is a real chicken casserole from the collection, of the 4,042. When we group recipes by name, ignoring words like easy, best and Grandma's, it is the dish people have written down more than any other, and almost all of them gave it the same plain name: the 4,042 recipes use only different spellings.

The recipe Americans write down most is chicken casserole

There are 4,042 different chicken casserole recipes in the collection, about 1,250 more than the dish in second place, and the next two places also go to casseroles.

The list below is the top 20 dishes by how many versions exist. It reads like the table at a church supper: broccoli casserole, cheese ball, meat loaf, chili, fruit salad, baked beans, pecan pie, cranberry salad. There is nothing from a restaurant menu and nothing you would need to look up, because these are the recipes people kept for themselves and passed on.

The grey column on the right explains it: for most of these dishes, nine out of ten versions come from cookbooks.com, the site built on typed-up family and community cookbooks, so this is the menu of American home cooks writing for each other. Banana bread is the most modern dish on the list, and even there seven versions in ten come from the old cookbooks.

A cornflower CorningWare casserole dish with a golden crumb top
different chicken casserole recipes.

of the 4,042 chicken casseroles are different from each other. The single most common version, chicken, cream of mushroom soup, onion soup mix and rice, appears only four times.

So the most popular dish is also one that almost nobody makes the same way. Most cooks start from chicken and a can of creamy soup baked until it bubbles, then add their own topping or vegetable, and that small change was enough for them to write it down as their own.

What goes into a chicken casserole

With thousands of versions of a dish, you can see what most people put in it. The bars show, out of every 100 chicken casseroles, how many use each ingredient.

58 in 100 list plain chicken and 34 ask for chicken breasts, so the meat is written down in different ways, but it is there in nearly every version.

The sauce comes from a can: 56 in 100 use condensed cream of chicken soup and 39 use cream of mushroom, often both together, so nobody has to make a sauce from scratch.

Meat loaf

Ground beef, eggs and onion are in almost every meat loaf, and four in ten put ketchup in it or on top. What changes is the filler: a third use breadcrumbs, and the single most common exact version uses oats instead.

Chili

A typical chili needs nine ingredients. Ground beef, onion and chili powder are in about four versions out of five, and then it splits: about a third add kidney beans, and the rest leave them out or use another kind.

Pecan pie

Pecan pie is almost a fixed formula: eggs in 99 versions out of 100, butter in 83, then vanilla, sugar and the pecans themselves. The 28 in 100 that use flour are probably counting the crust, and about one in six simply asks for a ready-made pie shell.

Banana bread

Banana bread is a baking recipe, so it behaves like one: eggs, flour, sugar and baking soda in nearly every version. Only 77 in 100 list banana under that exact name, because some write "ripe bananas" or "mashed bananas", which our counting keeps separate.

A few ingredients turn up everywhere

Across all two million recipes, a small group of ingredients does most of the work, and the first few are in a quarter to a half of everything.

As you scroll, each square below stands for one recipe out of a hundred, and the dark ones are the recipes that use the ingredient.

Salt
46
in every 100 recipes

Salt, 46 in 100

Salt is in almost half of all recipes. That also means more than half never mention it, partly because so many are sweet, and partly because plenty of cooks take salt for granted the way they take water for granted.

Butter, 30 in 100

Butter is in nearly a third of all recipes, in cakes and cookies but also in a lot of casseroles, vegetables and sauces.

Eggs, 28 in 100

Eggs are in 28 recipes out of every 100. They hold cakes, breads, meat loaves and casseroles together.

Sugar, 28 in 100

Sugar is level with eggs, and the two turn up together nearly twice as often as chance would explain, because between them they are the start of most baking.

Onion, 22 in 100

Onion is the first ingredient on the list that no cake would use, and it is in 22 recipes out of every 100.

Flour, 22 in 100

Flour is level with onion. Most of it goes into baking, and some of it is a spoonful used to thicken a gravy or coat a piece of chicken.

Garlic, 20 in 100

Garlic is in one recipe in five overall, but that average hides a big change between old and new recipes, which we come back to further down.

After those seven, the numbers drop quickly. Black pepper is in 18 recipes out of every 100, milk in 16, olive oil in 13 and vanilla in 12, and by the 30th ingredient on the list you are down to three or four recipes in every hundred.

Behind them sit about 170,000 other ingredient names, and most of those are used by only a handful of recipes: a particular brand of hot sauce, a jam somebody made from their own plums, a cheese from one region. That long tail matters later on, because it is the reason so many recipes have one ingredient you don't have.

What goes with what

Some ingredients almost always arrive together. Pick one to see, out of every 100 recipes that use it, how many also use each of its most common partners.

Salt sits near the top for nearly everything, so look at the second and third places, which is where the real partnerships show up: olive oil and garlic, butter and sugar, tomato and onion.

The closest pairs

Common pairs are not always close ones, because salt and eggs turn up together all the time simply because both are everywhere. So we also measured how much more often two ingredients appear together than they would if cooks picked ingredients at random. The pairs below are the tightest among the 150 most common ingredients, and they read like a map of cuisines: sesame oil and sesame seeds, coriander and turmeric, cilantro and lime, and from the old cookbooks, pineapple and marshmallows.

"26×" means the two appear together 26 times as often as chance would predict. Only pairs found together in at least 2,000 recipes are included.

Most recipes need about eight things

Count the ingredients in each recipe, leaving out water, and the most common answer is seven or eight.

About one recipe in four gets by with five or fewer, and only one in twenty needs fifteen or more, so the long, intimidating recipe is much rarer than it feels when you are reading one on a phone in the supermarket.

Number of ingredients in a recipe, and how many recipes need that many. The green columns are seven and eight.

The top dishes sit right in the middle. A typical cheese ball or fruit salad needs six things, most casseroles need seven, and chili, lasagna, beef stroganoff and potato salad need nine or ten. Even lasagna, the longest of them, needs only ten.

Grandma's kitchen and today's

The collection holds two very different kinds of kitchen, and the easiest way to see them is to compare where the recipes came from.

On one side are 787,461 recipes from cookbooks.com, typed up in the shorthand of old family and church cookbooks, where 81 out of every 100 write a cup as "c." On the other are 299,031 recipes from today's food sites, Epicurious, MyRecipes, Food Network, Food52 and Serious Eats, where only about one in a hundred writes it that way. The collection does not record the year each recipe was written, so we can't date them, but the two kitchens cook very differently.

Grandma's kitchen
recipes
Today's food sites
recipes

The can of soup

One old recipe in 23 uses a can of cream of mushroom soup. On today's sites it is one in 501, about twenty times rarer.

Cool Whip

Cool Whip is named in one old recipe in 23, about as often as the soup, in pies, salads and anything layered in a dish. Today it is one in 474.

Jell-O

Jell-O is in one old recipe in 30, often in the "salads" that were really desserts. On today's sites it is one in 595.

Garlic

Garlic was in one old recipe in 13 and is now in one in three, roughly four times as often. Garlic powder fell from one old recipe in 50 to one in 667, so fresh garlic took its place.

Olive oil

Olive oil went from one old recipe in 48 to one in four today, about fourteen times as often.

Cilantro and lime

Cilantro went from one recipe in 340 to one in 16, and lime from one in 181 to one in 17. More than half of all old recipes ask for something by the can, box, packet or jar, against about one in five today.

Close-up of a cream-coloured oven temperature dial

Half of all ovens are set to 350

731,925 recipes tell you an oven temperature, and half of them say 350°F. The next most common, 375°F, is used by about one recipe in eight.

Before ovens had thermostats, recipes asked for a slow, moderate or quick oven and left the cook to judge it. When ovens with numbered dials arrived in American homes in the 1920s and 30s, a moderate oven became 325 to 375, and 350 sits right in the middle. Those three settings still cover 71 out of every 100 recipes that name a temperature, so recipes mostly still say "moderate oven", only as a number.

It depends on what you're baking

We split the recipes by the kind of dish in their name. Brownies, meat loaf and casseroles are the most loyal to 350, at about three or four recipes in every five. Muffins most often go to 400 and biscuits to 450, and only about three pizzas or roasts in ten use 350, and most of the rest go hotter.

Share of recipes of each kind that use 350°F as their first oven temperature.

How far your kitchen gets you

A recipe only works if you have every ingredient on its list. So we asked: if you stocked a kitchen with the most useful ingredients first, how many recipes could you cook?

We started with an empty shelf and kept adding whichever ingredient finished the most recipes. Sugar, butter, salt, flour and eggs come first, because baking recipes share so many ingredients, and onion does not arrive until the 38th place.

Even a kitchen with 300 different ingredients can make only about one recipe in four, and another 662,164 recipes would be just one ingredient short. The missing thing is different almost every time: the most common one, unsweetened chocolate, is behind only 1,549 of them. The rest are spread across thousands of other ingredients.

That is why the finder below shows you two things: what you can cook now, and what a single extra ingredient would unlock.

What can I cook?

Add what you have in your kitchen. The finder searches 753,765 recipes from twelve recipe websites that are still online and shows the ones you can make, plus the ones you are one ingredient away from. Each result links to the recipe on its own site.

Nothing added yet. Try "Add the basics", then add one or two things from your fridge.

The finder knows the 1,000 most common ingredients. It treats water as something you always have and does not know about quantities, so check the recipe before you start.

Get the next one free

The Data Drop sends one data story like this most Tuesdays. It's free, and you can unsubscribe whenever you like.

How we did it

The recipes come from RecipeNLG (Bień and others, 2020), 2,231,142 recipes gathered from recipe websites, each with a list of ingredient names that the researchers pulled from the original ingredient lines. We removed recipes with the same name and the same ingredients, which left 2,089,091. This page publishes counts, and the finder shows only each recipe's name and a link to the site it came from.

We tidied the ingredient names by removing words like chopped, fresh and large, turning plurals into singulars and merging common duplicates, so unsalted butter counts as butter and catsup counts as ketchup. Some things a cook would call the same are still counted separately, like chicken and chicken breast, or banana and mashed bananas. Water is left out. Dish names were grouped after removing words like easy, best, homemade and Grandma's. Brand names were matched in the original ingredient text.

Oven temperatures are numbers between 200 and 550 followed by a degree sign or the word degrees in the method. "The closest pairs" compares how often two ingredients appear together with how often they would if they were independent. "How far your kitchen gets you" adds ingredients one at a time, always picking the one that completes the most recipes, and counts a recipe only when every ingredient is on the shelf.

The finder covers recipes with up to 16 ingredients from Food.com, Allrecipes, Epicurious, MyRecipes, Cookpad, Taste of Home, Food Network, Food52, Food & Wine, Serious Eats, Delish and Land O'Lakes, because the other sources in the collection, including cookbooks.com, are no longer online. The collection leans heavily on American websites, so all of this describes American home cooking as it is written online.

  • Bień M, Gilski M, Maciejewska M, Taisner W, Wiśniewski D, Ławrynowicz A. RecipeNLG: A Cooking Recipes Dataset for Semi-Structured Text Generation. INLG 2020.
  • Eat Your Books, Why is 350 degrees such a common oven temperature? (2021).

sheets.works is not affiliated with any website, brand or company mentioned.