First, keep in mind that place of origin varied depending on the port the slaves were imported to. DNA testing has revolutionized our understanding of African American ancestry. On average, an African American is over 20 percent European and 1 to 2 percent Native American, and virtually no African Americans have fully African ancestry. A third of African-American males are directly descended from a white male.

As this study shows, there are still significant differences between local populations in the United States. African Americans from South Carolina are much more likely to be descended from Senegalese ancestors than African Americans from Virginia, who in turn are much more likely to show signs of Nigerian ancestry.

This map and break-down of geographic origin of imported slaves is rather accurate Most European trading posts were on the coast because it was almost impossible for Europeans to penetrate the interior. Besides the obvious of jungles and malaria there was a disease that killed horses and the rivers were very difficult to navigate. A lot more going to the Americas came from the west coast because there were more posts, more European control over trade and it's just a shorter trip.

Here's another map It's very similar but you might be interested in the chart at the bottom.

While I am busy compiling a more erudite reply to add to what others have already said, I thought I'd chuck you a link to the PBS series African American Lives. It's fascinating, and goes a long way towards deconstructing monolithic "black-white" American notions of race as it traces the profoundly complex ancestries of its interlocutors.

Warning: watching the series (which is also available via Netflix and/or YouTube, I believe) may lead to joining ancestry.com or other such behaviors.

Edit to add an actual reply:

So very smart folks have already mentioned the PLOS article and have given you really good maps that help understand where different populations came from and ended up. Another interesting way of investigating this question is to look into creolized languages and maroon populations.

For instance, what does one make of a language spoken in Jamaica that bears a strong resemblance to Ashanti? Well, you can trace that linguistic difference back to the Ghanaian Coast. (Sorry this link is so bad--if anyone has a better, publicly accessible source on the Coromantee, that'd be welcome.)

A creole language also helps us isolate the Gullah and their connection back to Africa, especially to Sierra Leone.

If you're curious, this article provides more of an introduction to creole languages and how we can use them to understand more about the African diaspora and its movements. A few pages in, you can find some neat maps of different creole languages in the New World.