People talk about the effect of official lockdowns vs. citizen decisions on the transmission rate of the #coronavirus . I really like the combination of these 2 graphs:

Here we can see the effect that a delay of 5 days had on the curve of NY cases vs. Spain [1/7]

NY was trailing Spain pretty closely in # of cases. But Spain ordered its lockdown on March 15. Two weeks later, cases were going down. NY took 5 more days, so its # of cases blew past Spain's. As a result, so far Spain has had 235k cases vs 360k for NY. [2/7]

But we can also see that NY's curve started bending before the 2 weeks since the lockdown. Why? Because of changes in citizen behavior. [3/7]

By the time the shelter-in-place was effective in NY, on March 22nd, ppl had already stopped asking Apple for directions. It's likely they had curtailed some of their movement already, especially superfluous trips. [4/ 7]

We can see that effect on the NY case curve, which starts bending before 2w have passed since the March 22nd shelter-in-place. But it doesn't fully bend until the order was active.

That suggests that BOTH citizen behavior AND gov orders had an impact on #coronavirus spread [5/7]

Side note: movement in both countries did not change at all until March 10th, despite raging outbreaks in Iran and Italy. However, starting on March 10th—the day we published Coronavirus: Why You Must Act Now—mobility starts dropping. [6/7] https://medium.com/@tomaspueyo/coronavirus-act-today-or-people-will-die-f4d3d9cd99ca

This can be seen across countries: SK and IT had outbreaks earlier, so citizens lowered their mobility earlier. But other Wester countries remained unaffected until March 10th. Correlation is not causation, so maybe it was just incredible timing for the article. [7/7]