Hawaii is terrible at social distancing, it turns out.
Mobility data collected using cell phone location information — including cell site data and GPS or bluetooth beacons — shows just how much we’ve cut loose in recent months.
The data also gives us clues about where people spend their money, which could be used to plan for economic recovery.
In March, when the governor issued an emergency proclamation and everything shut down, the streets emptied out as most people diligently stayed in their homes as they were supposed to.
“As a whole, the state decreased its mobility uniformly,” said Kendrick Leong, a research specialist with the Hawaii Data Collaborative. “It was a testament to how united we are, how willing we are to make sacrifices for the good of the community.”
Then slowly, as the number of new and active COVID-19 cases began to dwindle, restrictions started to ease with parks, beaches, restaurants and bars reopening.
People were moving around more, and that showed in the data. Cases started to creep back up as restrictions loosened and began to spike in late June and early July.
And here we are now, seeing hundreds of new cases daily, with some hospitals nearing capacity.
As a result, state and county officials are bringing back sweeping restrictions, including a quarantine requirement for people traveling interisland, as well as creating a special Honolulu police unit to catch people violating restrictions on outdoor activities.
Carl Bonham, executive director of the University of Hawaii Economic Research Organization, told Civil Beat’s editorial board that the state needs a stronger communications plan because the public is not getting the message that it needs to take personal responsibility for preventing the virus from spreading.
“It’s unclear that the people who are behaving badly actually know what they’re doing to their community and to their job prospects and the future of Hawaii by leading to the next outbreak,” he said.
So how badly are people behaving?
Hawaii earns a solid D-minus in social distancing, according to one scoreboard from a company called Unacast. That’s decidedly a failing grade by any educational standard, but it’s slightly better than the U.S. national grade of F.
There’s almost no point in measuring where we rank nationally because everyone is doing poorly.
Data providers like Unacast and many tech companies, including universally known ones like Google and Apple, took to mining sources of different data to aid in the pandemic response, and that included cell phone location data.
That data, which is aggregated and anonymized in various ways to protect privacy, could show where and when people were moving to a certain extent, therefore helping us measure social distancing.
Government agencies are using this data to help inform their pandemic response plans and craft policies. For example, a company called SafeGraph provided its mobility data to the Centers for Disease Control and Prevention.
The state of Hawaii is not, though. The Department of Health said in an email statement that it was looking at sources of that information early on, but that it is no longer, without explaining why. But the statement also said, “Our investigations tell us people are gathering in droves.”
Unacast, which also provides a data dashboard on the retail impact of COVID-19, uses mobile phone data from apps and their interactions to calculate reduction in average mobility, nonessential visits and human encounters, according to its methodology.
Hawaii’s data from the company shows that immediately following the statewide stay-at-home order on March 25, the average mobility took a nosedive. From a 40% to 55% reduction on March 25, it dove down to a 70% reduction on March 29, earning the state an A grade.
The average mobility remained quite low until it began to creep back up starting in May, then began a generally upward journey with occasional dips.
Of the counties, Hawaii County fared best, earning a C-minus grade. Unacast’s data showed that it had the biggest decrease in human encounters, which makes sense, because it’s the least dense island in Hawaii.
To no one’s surprise, the City and County of Honolulu had the worst grade, at D-minus. But come on, we’re dense, the most heavily populated and have more traffic. Honolulu failed in two categories — nonessential visits and human encounters, while faring fairly well in reduction in average mobility.
Other providers’ data, including from Descartes Lab, which the Hawaii Data Collaborative works with, show a similar trend: In Hawaii, mobility decreased dramatically when the shelter-in-place order went into effect, then as restrictions began to ease up, it began to creep back up.
Data also suggests that social distancing may be loosening nationwide.
Unacast’s data shows that it has waned in recent months, says Tanya Merisier, a spokeswoman for Unacast. It appears that people are generally adhering to social distancing practices less and less, even as new hotspots are springing up across the country, especially in states like Florida, Texas and California.
“While the number of new daily cases grows, residents are eager to return to pre-COVID activities,” she said.
Mobility data’s only utility isn’t just to measure social distancing, Leong of the Hawaii Data Collaborative says. It can also be used as a proxy for economic activity.
“We can assume the places they are going and spending money,” he said.
This data can be used as a supplement to other data, including unemployment rates and job numbers, to help determine economic recovery policies and plans, he said.
Of course, mobility data has limitations, he said. Because of privacy concerns, it has to be aggregated or summed up to a fairly large extent, so no specific person can be identified.
That means granular comparisons are often impossible. For example, it would be useful to compare how mobility differs between the more affluent parts of Honolulu versus others where more essential work, such as construction work, is being done, Leong said.
But because most datasets don’t break down that finely, that’s not possible, he said.
Some other companies provide different details using other methodologies, though. For instance, Apple’s COVID-19 mobility data shows how driving trends have changed, because they use the data from people’s map apps. For Hawaii, it says driving decreased by 36% since January. Google’s report shows how mobility has changed in different areas, like parks versus retail and recreation.
No matter what any of these numbers say, one thing is for sure. “We may never come back to pre-pandemic mobility patterns,” Leong said.
COVID-19 has changed the definition of “normal” forever. Things will probably never be the same.
Civil Beat is a small nonprofit newsroom, and we’re committed to a paywall-free website and subscription-free content because we believe in journalism as a public service.
That’s why donations from readers like you are essential to our continued existence.
Help keep our journalism free for all readers by becoming a monthly member of Civil Beat today.