Springtime social distancing in North Carolina
Surprises in March and April
May 1, 2020
To explore the latest trends in social distancing, I conducted a quick analysis of mobility changes in North Carolina since the start of the quarantine period.
Methods. The data are based on location tracking apps on mobile phones, provided by CubeIQ via the University of Maryland. The key measurement here is the percent of residents staying at home (i.e., no trips more than one mile away from home). Sadly, we don’t have much more information than that on methods from the data source, so we have to take things at face value here. We calculated the county population weighted averages for this analysis. The fit on the figure in the left is done using LOWESS, a type of locally weighted polynomial regression that reduced noise from low mobility on weekends to reveal the broader pattern over weeks.
By the time the Executive Order 121 to stay-at-home came around on March 30th, North Carolinians had pretty much decided to stay put, figure on left. There wasn’t a substantial increase after EO 121. This suggests that people may voluntarily have limited their movements out of personal concern during the last 2 weeks of March. It also calls into question the emphasis on legal mandates by protestors advocating for loosening restrictions, and also is a warning that “normalization” may not mean a return to our previous lives. Schools and restaurants closing on March 16–17 did have an appreciable impact on staying at home.
Over the past 5 weeks, we also observed:
- Sundays are the most homebound day, with more than 40% of NC residents staying home.
- Fridays are the least stay-at-home days.
- Easter Sunday was the most homebound day observed.
- As the weather has gotten warmer (or quarantine fatigue has set in?), fewer people are staying at home.
Updated May 2, 2020 at 9:30 am ET:
Late last night, NPR published similar trend data showing similar patterns in TX and VA (as well as nationally), even with the same color scheme! Nationally, there has been a mobile devices staying home in the last 7 days.
And, in Virginia too, much of the stay-at-home behavior was observed before official order from the governor.
Limitations should be kept in mind. These data provide a snapshot of social distancing trends in NC, as measured through the proxy of a lack of movement from home. There are many other forms of social distancing that aren’t reflected here, like wearing masks. These data do not provide us with information on 6 feet apart spacing. Movement of less than a mile means different things in car-dependent rural areas versus in urban areas of our state. Those in the urban cores could have travelled less than a mile and visited many establishments. Or they could be social distancing hard and still traveling about: car data from friends Rebecca and Robby in Raleigh show them moving 3 miles/day… because that’s the only way to get their 2.5 year-old daughter to nap. So, these data aren’t perfect, the metric is a proxy, and we will take rurality/urbanicity into account in future analyses, but this is feature that hasn’t changed during this time period. Also, weather and holidays have an impact we haven’t documented yet, but can see anecdotally in the data.
