Fig. 4: Time series cross-correlations and optimal lags between respiratory virus transmissibility (time-varying effective reproduction numbers, Rt) and cell phone mobility in the greater Seattle region, September 2019 – May 2020.
Points are individual mobility indicators derived from aggregated mobile device location data. Spearman’s rank correlation coefficients are on the y-axis, and temporal lags (in weeks) between \({R}_{t}\) and mobility are on the x-axis. Negative lags indicate behavior leads \({R}_{t}\), and positive lags indicate \({R}_{t}\) leads behavior. A lag of 0 indicates the two time series are in phase (i.e., synchronous). The yellow shaded panel in each facet includes mobility indicators that have a leading, positive relationship with transmission, and hence would be considered predictive of transmission. Figure. S10 shows cross-correlations for the full set of mobility indicators.
