Quantitative understanding of the process of knowledge creation is crucial for accelerating the advance of science. Recent years have witnessed a great effort to address this issue by studying the publication data of scientific journals, leading to a variety of surprising discoveries at both individual level and disciplinary level. However, before scientific journals appeared on a large scale and became the mainstream for publishing research results, there are also intellectual achievements that have changed the world, which have usually become classic and are now referred to as the great ideas of great people. So far, little is known about the general law of their birth. In this paper, we reference Wikipedia and academic history books to collect 2001 magnum opuses as representations of great ideas, covering nine disciplines. Using the year and place of publication of these magnum opuses, we show that the birth of great ideas is very concentrated in geography, and more concentrated than other human activities such as contemporary knowledge production. We construct a spatial-temporal bipartite network to study the similarity of output structures between different historical periods and discover the existence of a Great Transformation around the 1870s, which may be associated with the rise of the United States in academia. Finally, we re-rank cities and historical periods by employing an iterative approach to study cities’ leadership and historical periods’ prosperity.