Here's a fact: Nearly everyone in America loves Chinese food. Who among us doesn't have childhood memories of takeout dinners served in the distinctive white boxes, followed by a humorous reading of fortune cookies? And Chinese restaurants, with their lengthy menus and Lazy Susans, have been a ubiquitous part of the American landscape for decades.
MIT Technology Review tells us a bit more about how they did it:
They began by downloading all the recipes from a Chinese recipe website called Meishijie. This contained almost 8500 recipes based on nearly 3000 ingredients. They grouped the recipes according to their origin in one of 20 regions. Finally they created a food web consisting of the set of all recipes on the set of all ingredients. Where recipe contains an ingredient they draw a link between them. Since each recipe belongs to anyone regional cuisines these links can then also be categorised into cuisines. Counting these links shows how prevalent each ingredient is in each cuisine.
Cuisine, of course, isn't the only aspect of China in which there's a lot of regional variation. Here's a map showing the country's languages and dialects:
There remains a perception of China is that it's a giant land of sameness, a billion-man nation of people who think alike, talk alike, and eat alike. Instead, it's a nation cobbled together through over 50 centuries of invasion, war, consolidation, and treaty, one that has only really existed in its present form for a brief sliver of time. These regional differences -- of language and cuisine -- are truly vestiges of an earlier, less unified era.