The Lifeboat News
[ Message Archive | The Lifeboat News ]

    Some country coronavirus data Archived Message

    Posted by walter on May 9, 2020, 11:36 am

    (As of last night, from past worldometer data)

    A change to the format, but the gradients data in the last 3 cols is as before: ie each country gradient shown (whether cases or deaths) is determined simply by comparing the most recent six-day period with the preceding one. The idea being to indicate acceleration (gradient>1) or deceleration (gradient<1). (Acceleration is equivalent to exponential growth.)
    These are then multiplied together and square-rooted to give one overall measure/indicator of 'acceleration' for each country. This, the main outcome of interest, is the last column; it shows a composite measure of acceleration in the gradients of cases and deaths for each country.

    Brazil with a combined gradient of 1.5 and deaths in the thousands, may not be far behind the European countries in number deaths, remains a worry with its huge population and seemingly erratic government. Also a concern numerically (deaths in the thousands, and gradient of >1) are Russia, Peru, Mexico. The countries that were an early concern (European+US) still are; gradients not low enough to suggest daily numbers are falling fast enough to bring the death toll into the 'controlled' zone.
    BTW the French data has been all over the place - even worldometer complained - and can't be taken as accurate over any small time period. The UK data were adjusted to backspread the care home and home deaths that were added en bloc.


    As usual, this shows, for my selected group of countries, past trends of 'acceleration' in combined cases and deaths based on these gradients.
    As of last night, now only two countries of these eight are in 'exponential' increase: India (still not reporting that many deaths, hard to gauge - reporting may be erratic?) and Canada. Sweden has come back to borderline exponential, may be temporary.

    The last column is an overall gradient for the top 16 in terms of (from memory) no of cases. Hasn't got cleanly away from 'exponential'.

    Whether country or 'overall', generally these combined cases and deaths gradients need to come down towards maybe 0.5 before giving numerical comfort (if such a thing exists :-) ) that the spread of the coronavirus is being brought under control or is subsiding.

    Message Thread: