COVID-19 Lockdown Unravels the Complex Interplay between Environmental Conditions and Human Activity

Figure 3

Causal analysis for the time course of the 16 observables. (a) Partial correlation network: each node corresponds to a variable, and the color encodes the type of variable meteorology, energy, mobility, and partial correlation value. (b) Granger causality network: the arrows are oriented in the causal direction; the variables which have a causal impact on are highlighted with solid edges. (c) Bayesian state-space model: the top panel shows the data and a counterfactual prediction for the lockdown period. The middle plot shows the difference between observed data and counterfactual predictions. The bottom plot is the cumulative effect of the lockdown. (d) Probability of inclusion for the regressors: light-blue bars represent negative coefficients, while red bars represent positive coefficients.

(a)

(b)

(c)

(d)