Fig. 2: Connecting morphology to vulnerability.
From: Interplay between population density and mobility in determining the spread of epidemics in cities
a Normalized epidemic threshold \({\tilde{\lambda }}_{{{\mbox{c}}}}\) vs. population density hotspot concentration κ for cities in each of the chosen four countries (color code). The Spearman correlation coefficient is rs = −0.45 (p-value < 10−9), showing a moderate but statistically significant correlation. The solid line shows the result of fitting the data to a power-law function \({\tilde{\lambda }}_{c} \sim {\kappa }^{\beta }\), obtaining β ≈ − 0.25 via least squares regression. The shadowed region covers 68.2% confidence intervals for the estimated parameters. b Histogram for the impact of reshuffling the flows connecting hotspots among their non-hotspots neighboring areas on the epidemic threshold, measured by the ratio \({\tilde{\lambda }}_{\,{{{\rm c}}}}^{{{{\rm MOD}}}\,}/{\tilde{\lambda }}_{{{{\rm c}}}}\), where the numerator represents the recomputed threshold after reshuffling. The green-colored (red-colored) zone corresponds with a beneficial (detrimental) effect of such intervention translated into an increase (decrease) of the epidemic threshold. c Bar plot containing the count of cities within each category according to the outcome of the removal of flows connecting hotspots for each country here analyzed.
