Read on Twitter
I love this heatmap
It represents over 100,000 individual datapoints.
These are Florida COVID-19 cases, over time, by age bracket.
I published open-source code to make it: https://github.com/mbevand/florida-covid19-line-list-data
The recent case surge is driven by 20-24-year-old Floridians.
1/N
It represents over 100,000 individual datapoints.These are Florida COVID-19 cases, over time, by age bracket.
I published open-source code to make it: https://github.com/mbevand/florida-covid19-line-list-data
The recent case surge is driven by 20-24-year-old Floridians.
1/N
The data comes from @HealthyFla who provides an amazing dataset: age information for all 100k COVID-19 cases in https://open-fdoh.hub.arcgis.com/datasets/florida-covid19-case-line-data
Here is the heatmap in numeric form, except aggregated over 10-year age brackets and 7-day time periods:
2/N
Here is the heatmap in numeric form, except aggregated over 10-year age brackets and 7-day time periods:
2/N
The heatmap pixel intensity is a bit more granular and represents the number of COVID-19 cases reported per 5-year age bracket per 4-day time period.
Notice how the peak in early April consisted of cases spread rather uniformly over a large swath of adults.
3/N
Notice how the peak in early April consisted of cases spread rather uniformly over a large swath of adults.
3/N
But since early June, young Floridians around the age of 20-24 appear to be the primary drivers transmitting COVID-19 to the rest of the population.
They're more likely to socially hang out w/people their age, so pixel intensities decrease progressively toward older ages.
4/N
They're more likely to socially hang out w/people their age, so pixel intensities decrease progressively toward older ages.
4/N
Also, notice how today the number of cases (intensity of pixel) is higher across ALL AGE BRACKETS compared to the peak of early April.
(Well, except ages 90+ but the record for this age brackets will be beaten in the next few days if current trends continue)
5/N
(Well, except ages 90+ but the record for this age brackets will be beaten in the next few days if current trends continue)
5/N
So you may see Florida news reports minimizing the danger because the disease is less severe in young adults.
But looking at the data, we *should* be worried: if all age brackets see more cases, daily deaths are expected to surge to new records.
6/N
But looking at the data, we *should* be worried: if all age brackets see more cases, daily deaths are expected to surge to new records.
6/N
In fact, based on this data, I expect Florida deaths to surge to new records in the coming 2-4 weeks, because the mean time from infection-to-death is 23 days (see page 4 of https://static-content.springer.com/esm/art%3A10.1038%2Fs41586-020-2405-7/MediaObjects/41586_2020_2405_MOESM1_ESM.pdf)
7/N
7/N
Other interesting artifacts—I looked at Florida #covid19 cases with 1-year granularity, and:
- peak cases are exactly at age 21 (legal drinking age in FL)
- among the population there are relatively few reported cases of children under 18. The cutoff is really sharp at 18.
8/N
- peak cases are exactly at age 21 (legal drinking age in FL)
- among the population there are relatively few reported cases of children under 18. The cutoff is really sharp at 18.
8/N
