Figure 6From: Traveling heterogeneity in public transportationCharacterization of all bus trips. In (a) we show, for each small circle, the relation between the fraction delay in function of the α exponent for Fortaleza. As we see, when \(\alpha \to 2.0\) the average fraction delay, calculated by the Nadaraya–Watson (red dashed line), grows fast. In (b) the same is shown for Dublin. In (c) we show the probability distribution for the Gini for Fortaleza (in blue) and Dublin (in red). One can see that the PDFs appear Gaussians, despite a slight assimetry when compared to the statistical fit (in dashed lines). To confirm that hypothesis we apply the Kolmogorov–Smirnov test. It obtained p equals 0 for Fortaleza and Dublin. Therefore the hypothesis that the distribution is normal is rejected. The kurtosis and skewness [31] were also computed. The kurtosis for Fortaleza is 1.9493 and the skewness is 0.5485. For Dublin, the kurtosis is 1.450 and the skewness is −0.0282. These results indicates that, despite the fact that one visually appears the normal fit, the PDF of Fortaleza is more asymmetric than Dublin’s oneBack to article page