Figure 5
From: Randomizing bipartite networks: the case of the World Trade Web
Application of our method to the binary, undirected, bipartite World Trade Web in the year 1963 (left column) and 2000 (right column). Panels report up VS Qp (a, b) and dc VS Fc (c, d). Observed points are in blue; the black solid curves are BiCM-induced ensemble averages; the gray dashed curves indicate the ±1 standard deviation region; the gray dash-dotted curves indicate the ±2 standard deviations region. Colored areas represent the ensemble density of expected points (sampling 5000 matrices). Our null model seems to satisfactorily capture both trends. Panels (e, f) show the so-called “poverty trap”, i.e. the group of countries with lowest fitness8,9. Notice how all such countries lie within the ±2 standard deviation region (or immediately outside).
