Figure 3

Network position and timing of biggest hit for movie directors. (a) We report the impact S of Francis Ford Coppola illustrating the case when PageRank centrality peaks first (\(\tau = 5\)), followed by the impact, while the example of George Lucas illustrates the opposite behavior (\(\tau = +1\)), where a peak in the impact is followed by the peak in networking. The figure then shows a comparison between the groups of film directors for whom success peaks before (colored continuous lines) and after (colored dashed lines) their network peak. For movie directors, (b) their network position based on their PageRank and their success measured by their Q parameters, (c) their network position based on their degree and their success measured as their highest impact, the binned (15 bins) distributions of the two groups do not show any significant difference based on the Kolmogorov–Smirnov test (\(d_{\mathrm{degree}} = 0.02\), \(d_{\mathrm{PageRank}} = 0.02\), \(p <0.0001\)). (d) Shows the distribution of the shift parameter τ between the directors’ network centrality (PageRank) and impact time series, coloring the distribution corresponding to the original data by orange, and to the randomized data by orange (KS test \(d=0.27\), \(p < 0.01\))