So we actually have a good bit of information available on the subject, thanks to shipping manifests, knowledge of colonial outposts, and, later on, naval interdiction accounts. And luckily, David Eltis, Martin Halbert, and a fairly large team of historians and programmers have been compiling all that information and developing the Slave Voyage Database for some time now.

Note: When I give percentages, they are pre-1700/post 1700 (ex. 60%/40%). The 18th century saw the slave trade grow exponentially with the boom of sugar and rice plantations (and to a lesser degree tobacco and indigo), and as such the is usually divided into those period.

The short answer is that the trade shifted around over time and place, but the majority - 60%/38% - of slaver voyages departed from West-Central Africa (roughly modern Angola, Gabon, Zambia, and Democratic Republic of the Congo). But this is also pretty misleading, because only a small fraction of voyages ended in North America, less than 2%/14%. In addition, while the majority of voyages are leaving from West-Central Africa, not all of those departing originated there.

One of the major misunderstandings that gets touted so often is that Africans were enslaving and selling their own people. While this is true in some cases, it is mainly a product of a western worldview in which Africans (as well as Arabs and Native Americans) are proudly classified into one homogeneous group. The reality is that just like Europe, Africa was home to many diverse populations, tribes, and kingdoms which regularly waged war against each other. These conflicts generally saw the victor enslaving at least some of the defeated population, though this enslavement usually differed from the chattel slavery that the Trans-Atlantic Slave Trade promoted.

So, with Europeans suddenly creating a major market demand for slaves, and in return offering important items (such as firearms, iron, rum) these coastal kingdoms were faced with a predicament. If they did not produce enough slaves for market, they faced the possibility of their own people being enslaved. One option was to sentence criminals to market, which happened though it never satisfied demands. The other was to wage war and slave raids against those on their borders, and sell those slaves to the Euopeans - in turn purchasing the weapons needed to conduct larger raids farther from home. In short, the system became self-perpetuating.

Because the locations of these raids are rarely documented, we have no real way of knowing the origins of those enslaved people who entered the Trans-Atlantic trade. Unless there is some information passed down through generations, such as oral traditions or languages - or in a more modern time DNA evidence which is not 100% accurate.

And, if all of this did not pose a large enough problem to pinpointing providence, we have to deal with the size of slave voyages. As the capacity of slavers increased, slave forts were not always able to fill these ships to capacity. In turn, a slaver may take months to fill up before making the voyage to the Western Hemisphere - hopping port to port and taking on victims as they were able. Once loaded on board, most voyages assigned slaves numbers and they are only referred to as such for the rest of the voyage and into their sale. As you can imagine, that makes identifying a particular person, on a particular voyage, at a particular port of embarkation extremely difficult. So in sum, we know that a large portion of African slaves entered the Atlantic world from West-Central Africa, but can't know for certain if that region was the one which provided for the relatively small North American slave market.

For additional reading, I highly suggest David Eltis' work for the statistics of it all. For an easier read, albeit with its own issues, see The Slave Ship by Marcus Rediker. Also I highly suggest Sowande Mustakeem's recent work Slavery at Sea which focuses heavily on the female experience of the slave trade. Also, the National Par Service has an ethnography report available that draws from Eltis and other works.