Recent NC Movement Changes
Update through May 5, 2020
This past week saw an incremental increase in the movement across North Carolina based on anonymized GPS location information from 1.7 million mobile devices. The increase is noticeable but not severe, and mirrors trends in other states. But there is no doubt that movement has been increasing over the past 2 weeks.
In particular, movement spiked suddenly on Friday May 1st. Friday is generally the day of the week with the most movement in North Carolina, but May 1st was way more than usual. So we investigated which counties were leading the movement on May 1st.
A handful of counties (Gates, Caswell, Jones, Montgomery, Duplin) had residents moving more than 15km on May 1st, increasing the average for the whole state. Scattered around the state, these counties do not show a discernible pattern, but perhaps local event knowledge could help explain it. The urban centers tended to have less absolute movement, which is in part a function of the density of the built environment. Gates County has had only 10 confirmed cases and no deaths. Caswell has had 38 confirmed cases and no deaths. Jones County has had 19 confirmed cases and 2 deaths. The virus seems not to be particularly prevalent in this area, insofar as we can tell by looking at cases-among-tested statistics.
So we then explored if movement has been changing over time in these counties. Relative changes in movement would take rurality out of the equation. We looked at week-over-week comparisons for the most recent date we had data for; we compared Tuesdays 05-May against 28-April. (Tuesdays are usually normal days for movement in North Carolina so they serve as a stable comparison for weekly time trends.)
To our surprise, 3-out-4 counties showed increases in mobility week-over-week. Notably, the most populous county, Mecklenburg, where Charlotte is located, showed a 4.4% increase in mobility week-over-week. A quarter of counties showed a decline in mobility too.
In summary, movement continues to increase in North Carolina. While these data cannot say if these represent flouting of the stay-at-home order, the trend towards more mobility suggests that behaviors are certainly changing. Gates County in particular seems to have a lot of movement, and it is rapidly increasing in the last few days.
Caveats
We have been using smartphone movement as a proxy for the stay-at-home component of social distancing, but there are limitations to this interpretation. Follow the link for full methodology on how mobility distances are calculated.
