Figure 1 (top) shows the temporal evolution of the entropy of the hashtags used by the followers of different media. This implies that even the first ranks are occupied by many different hashtags (notice that \(d(r) = 0.5\) corresponds to 180 different hashtags in position r; see “Methods”). Due to the large overlap of followers of different media, illustrated in Fig. 6, it is not surprising that the self-similarity curves of non-exclusive followers of different media show a qualitatively similar behaviour. Figure 7 Dynamics of self and cross-similarities corresponding to sub-populations that follow different media accounts in Twitter.

Source: New York Times March 07, 2023 16:55 UTC