COVID-19 stats
What is all this?
The purpose of this website is to track the development of COVID-19
pandemic worldwide and in particular countries in order to have a better
idea of what's going on and what to expect. There are many other trackers
with different focuses: here we focus on understanding the dynamics of
the infection numbers in different countries.
The website is updated daily and is always showing yesterday's data (we
think that it's good enough at this point).
The list below explains in more detail what each graph and parameter
means and how things are calculated.
- New cases
- The number of officialy reported infected people based on JHU data
(see below). Everyone, who has been officially diagnosed with COVID-19
will be counted in this number. It's almost certainly below the actual
number of infected people because of delays in reporting and not everyone
being tested, but this is the best basic data that we have.
- Deaths
- The number of officialy reported deaths from COVID-19 based on JHU
data.
- Recovered
- The number of people who have recovered from the disease. Some part
of this is estimated (at least in some countries) because people with
light symptoms are not hospitalized and their condition is not always
tracked.
- Active cases
- The difference between New cases and the sum of Deaths
and Recovered. This is the number of people that are currently
sick.
- X per day
- The change rate of X: the difference between todays number and the
number of the previous day. This is a discrete equivalent of
derivative of X.
- X per day change
- The change in the rate of change of X: the difference between todays
rate of change and yesterday's rate of change. This is similar to 2nd
derivative of X.
- Implied R
- Transmission rate, calculated based on the New cases per day
of the last two weeks. We take a weighted average of the new cases in the
last days (centered between 5 and 6 days ago) and divide the today's new
cases by that number. It shows how many new people were infected by one
person, who was officially counted as infected about a week ago. When R
is above 1, the number of infected people grows exponentially, when it's
below 1, it decays.
-
- Period
- The period control allows you to see all data or only last month,
last two weeks or last week.
- Smoothing
- The data as is contains lots of jumps. Some of these jumps are real,
some are artifacts of reporting. By averaging the numbers over time we
can get a less jittery curve where the main trends are easier to see.
Higher levels of smoothing result in more averaging. Note:
Smoothing is always applied to the main statistic and the rate of change,
rate of change of rate of change and R are calculated from smoothed data.
- SmoothType
- There are different ways to average over time: SMA is a simple
moving average, where the numbers of previous L days are added to today's
number and divided by L+1 (where L is the smoothing level); EMA is
exponential moving average, that starts with the original number and then
each day divides yesterday's EMA by (L+1)/L and adds today's number
divided by L+1; FWD repeatedly averages each day with the average
of the day before it and the day after it (except for the first and the
last) — it uses future data, so it's less backwards-looking but it
also overweighs the last data point.
- Stacked
- The graphs can stack countries data on top of each other. This makes
it easier to see the total amount (over all countries) and how each
country relates to the total. We don't stack the change in the rate of
change because these numbers are often negative and adding them looks
confusing. We also don't stack the implied R because transmission rates
are not additive (so R of the whole is not equal to the sum of R's of the
parts).
More information