New analysis from Pew Research offers a fascinating look at a population that’s central to the American political and cultural conversation: immigrants in the United States illegally.

Pew has calculated the size of the undocumented population for years, offering a well-founded estimate that there are about 11 million people who fit that description in the country. That figure has been fairly flat for several years now; in fact, in the wake of the recession, the number of people leaving the United States to go to Mexico was higher than the number immigrating illegally in the other direction.

As you might expect, the majority of undocumented immigrants live in large urban areas. About 6-in-10 live in 20 metro regions, Pew estimates. The cities with the largest populations of immigrants here illegally are also the largest cities: New York and Los Angeles. (All estimates are based on 2014 data.)

While there’s correlation between immigrant populations, big cities and Democratic politics, there are plenty of red areas that are home to undocumented populations as well. Looking at the House districts at the center of the metro areas, more red pops up on this map than you might expect.

As a percentage of the total population, though, the picture is different. There are isolated pockets which, in Pew’s estimation, have a disproportionately large percentage of the total population that immigrated here illegally.

Notice California, though. A large percentage of the population of much of the state’s Central Valley is made up of people without documentation. On Thursday, the New York Times looked at the looming crisis facing that region’s farmers. Many voted for President Trump, but many also employ people who are at increased risk of deportation.

Below, an interactive version of the map that allows you to explore the estimated populations and density within the total population of undocumented immigrants. (Populations listed as 2,500 are categorized as 5,000 or fewer by Pew.) The picture of America’s immigration population may be more complicated than we assumed.