CV Leaderboard

Each metric is described below but first a word of caution...
You will see some very high numbers in places, say, a fatality rate of 10.0%.
This is not due to incorrect data. It's normal, early in a disease lifecycle, for numbers to be unrepresentative for example when there's only 10 (known) cases and one fatality
we (correctly) report a 10% fatality rate but of course this isn't representative of the entire area and population.
As more people get tested, we'd expect numbers to adjust into more normal ranges.
Our point is the numbers here are accurate and correct, but **for today only** - they are not a forecast of what is to come.

Doubling time is just the time it takes for a population to double in size/value. We use a 4-day lookback to calculate doubling time for example let's say on Monday there's 1,000 cases. 4 days later, on Friday, there's 1,150 cases. What's the doubling rate? In Excel, the formula is:

=(4*LN(2))/(LN(1150/1000))

Which outputs **19.8 days** as our doubling rate.
We do not know if there is an international standard for how many lookback days are best used in a pandemic.
More background at Wikipedia at the link here.

There are many good discussions elsewhere on case fatality rate and infection fatality rate.
This site uses "fatality rate" which is simply: (number of deaths) ÷ (number of cases).
For example if there are 2 deaths and 100 cases in a region, we show a fatality rate of 2.0%.
We understand this is not true CFR nor IFR, which is why we do not use those terms.
Due to low testing, fatality rates may be extremely high in places - like 9.0% - impossibly high **for now**.
This is simply because testing may not be widely available in the region or people are not getting tested.
This does not mean the **current** data is bad or wrong. It is correct! But it is not a forecast of what the future holds - it is just the data we have **now**.
In time, with more testing, the numbers will adjust into more expected ranges.

This metric (also known as *per capita*) uses the number of coronavirus cases and the area's population. For example let's say in the state of Connecticut
there are 34,855 cases today with 3,565,287 living there per Census.gov.
In Excel, the formula is:

=(34855*100000)/3565287

Which shows us Connecticut would have 977.62 cases per 100,000 people today.

This metric (which would be "Crude Death Rate" if we were doing fatalities per 1,000 people) uses average number of daily fatalities (last 7 days in) normalized for the area's population. For example let's say the state of Colorado for the most recent 7 days reports these fatality numbers: 1, -2, 0, 17, 1, 2 and 11 (yes, many states and counties report negative numbers due to corrections). Our average is 4.29. There are 5,758,736 people living in Colorado per most recent information at Census.gov. In Excel, the formula is:

=(4.29/5,758,736) × 10,000,000

Which shows us Colorado has 7.4 fatalities per 10,000,000 people today. For more background see the World Health Organization at the link here or WikiPedia at the link here.

Gray bars on some of the charts mean the number of cases in the region are small, less than 100 cases. We color them gray so people are not alarmed because often their metrics are extreme due to the small case counts.

Same as *cases per 100,000 people*, just use the number of deaths instead of case counts.

Each county is compared to a nearby county. We determine which counties are "nearby" by calculating the distance from the center of the county (as determined by Google Maps) to the center of all the other counties.

Coronavirus data currently sourced exclusively from New York Times's database

© 2020 CV Leaderboard