Tropical Africa presents many of the same problems as lowland South America to the investigator of the past.* The environment precludes easy excavation in many areas (especially in the West African rainforests, which are famously dense), and even in areas where it doesn't, many of the materials used in indigenous African cities were organic and thus inherently less lasting than hard rock. Some African capitals were moved frequently, as was the case in Buganda, one of the best-documented kingdoms in Africa. The capital was described by Europeans and Americans (including Henry Morton Stanley), many of whom were in awe of its massive wooden buildings and well-designed layout, but the city moved - dismantled and reassembled elsewhere - fairly often, and the building materials were all biodegradeable. If you're wondering why the numerous kingdoms of tropical Africa have failed to leave tangible, beautiful remains and extensive UNESCO World Heritage sites, it isn't because the people of Africa were not industrious or didn't build anything. It's just that they seldom built in stone, they liked to move about, and they often lived in environments that degraded or concealed the organic materials they used.
There are exceptions. The site of Great Zimbabwe, likely built by Shona-speaking people between the eleventh and fifteenth centuries, consists of elegantly-constructed concentric stone walls. The walls were formed of granite - or of whatever rock was to hand at other zimbabwes (the name means something like 'court', and there are other zimbabwes in addition to Great Zimbabwe). They were filled with debris to reduce the labour needed and tapered to ensure the structural integrity of the site. Only the walls remain at many zimbabwes, but actually the bulk of the construction was in daga, a kind of mud. Daga-built domed-roof huts connected by semi-randomly-arranged tapered granite walls - that's what zimbabwes once consisted of. The daga has since eroded, and only the walls remain.
The zimbabwes were probably ritual sites, or at least they were purpose-built for an elite. They appear to have served little use beyond enhancing status. They displayed power - in an obvious, overwhelming way - both to the ordinary people living in the society that built them and also to archaeologists today. The ordinary people of Zimbabwe society would have lived on a fairly small plateau, an island of productive land in a country dominated by the tsetse fly, a fly with the malign ability to slay both person and beast with disease. It is particularly bad for cattle, and hence also for pastoralists. The tsetse-free plateau provided good, productive land for cattle, but, of course, that land was at a premium. Combined with sorghum agriculture, the plateau generated a surplus of food, and it also harboured large gold reefs and iron deposits, enthusiastically mined by Great Zimbabwe's inhabitants. As in so many other places, the population pressure that resulted from the desire for good land, and the land's ability to sustain a large population, generated elite power through control and delegation of land and food. The stone walls, some of the most impressive sites of prehistoric tropical Africa, are the most visible signs of this elite power.
Ever since the ruins were first seen and documented by Europeans - probably first investigated by Carl Mauch in 1871 - they have been ascribed to the work of Phoenicians, or lost tribes of Israel, or white people, or Atlanteans, or anyone but Africans. The whole issue has become highly politicised, especially in a land of very obvious racial and economic divisions, and especially in a land named after the site under discussion (Zim is named after the prehistoric buildings). Racists point to the lack of stone architecture in the rest of Africa, except the east coast (where the racists, of course, ascribe it entirely to Arab influence). Foreign objects actually have been found in Great Zimbabwe, including small amounts of Chinese porcelain, indicating some minor degree of influence from outside the immediate area. But the archaeological evidence is pretty clear: the zimbabwes, including Great Zimbabwe, were built by African people, and the only reason other enormous signs of social complexity are absent from Africa is that most other African groups rather sensibly built in wood and leaf.
Africa has always been sparsely-populated, at least compared to Eurasia and the Americas. Tsetse flies and malaria have had much to do with that. But African civilisations achieved high levels of complexity before European imperialists arrived, and neither urbanisation nor complex 'civilisation' were absent in African prehistory. Iron was worked relatively early and had spread south of the Congo by the eighth century CE. Cities grew up in many areas before Europeans or Arabs arrived on the scene. Despite presumptions to the contrary - that African cities only rose where foreign influence was important, that African societies failed to develop complex societies without Eurasian tuition, that white people must have been responsible for Great Zimbabwe or the cities of the Sahel - archaeology is increasingly showing earlier native precedents for the complex societies that Arabs and Europeans encountered and described on the continent. In the Sahel, the African rainforest, the now-largely-Muslim east coast, in Zimbabwe, around the East African Lakes: everywhere social complexity appears to have begun before or during the first millennium CE, and it appears to have resulted from pressure on land and resources in almost every case.
Great Zimbabwe probably declined in the fifteenth century. The Arabs and Portuguese of the sixteenth century describe it as more-or-less ruined. The reasons for the collapse are not hard to discern: As the tsetse free land was so desirable, no one would have wanted to leave it. The land became overgrazed, and disease would probably have been rife - people in medieval Zimbabwe did not have much incentive to travel very far to defecate. The area would also have been denuded of trees for cooking fires, and the fires themselves, needed as they were by a large urban population, would have created a layer of smog on the plateau. After four hundred years of habitation and trampling by people and cattle, the desirable aspects of living there would have been easily countered by environmental devastation and the lack of sanitation.
*This is only true of archaeology and genetics. The linguistic evidence is pretty clear: the Niger-Congo language family, including the Bantu languages of southern Africa, appears to have spread out from West Africa - today the densest-populated part of the continent - about two thousand years ago, probably as part of a cultural and genetic complex consisting of language, pastoralism, iron-working, and yam agriculture. Confirming this archaeologically is tricky, as much of the evidence has inevitably disappeared in the tropical landscape and dense rainforest. The slight evidence that does exist correlates well with the impression from linguistics, however.