Read on Twitter
this is simply wrong.
of those 120 deaths, only 8 were actually from 7/9. the rest were backfill to earlier dates.
this is the real death on date chart for florida. (courtesy of @JoshFriedlander )
the total for all of july is 135 so far.
120 did not happen yesterday. https://twitter.com/reporterannie/status/1281575230801088512
of those 120 deaths, only 8 were actually from 7/9. the rest were backfill to earlier dates.
this is the real death on date chart for florida. (courtesy of @JoshFriedlander )
the total for all of july is 135 so far.
120 did not happen yesterday. https://twitter.com/reporterannie/status/1281575230801088512
so let's take a breath here and learn to read the data.
data gets reported in boluses and most of the deaths are NOT on the day they get reported as.
because TCP and WoM suck, they are reported as day of report (top series) but what really happened is the bottom series.
data gets reported in boluses and most of the deaths are NOT on the day they get reported as.
because TCP and WoM suck, they are reported as day of report (top series) but what really happened is the bottom series.
so backfill from the past gets mistaken for a spike in the present.
it's a really bad way to present data if you're looking for trends. and ALL the US data is like this unless you find one of the rare sources that does real day of death (FL has one)
http://ww11.doh.state.fl.us/comm/_partners/covid19_report_archive/state_linelist_20200710.xlsx
it's a really bad way to present data if you're looking for trends. and ALL the US data is like this unless you find one of the rare sources that does real day of death (FL has one)
http://ww11.doh.state.fl.us/comm/_partners/covid19_report_archive/state_linelist_20200710.xlsx
and it's being used to spread hysteria when, in fact, it's just bad data handling.
the data quality from most US sources is lower than a snake's belly.
so be VERY careful with it. it's incredibly easy to misuse and misunderstand if you have not spent a lot of time with it.
the data quality from most US sources is lower than a snake's belly.
so be VERY careful with it. it's incredibly easy to misuse and misunderstand if you have not spent a lot of time with it.
now, let's address the other issue: day of death reporting lags. that 8 number from yesterday WILL rise as it backfills. it might double to 16 today as new reports come in. it may be a week or so before we can be 90% confident in it.
so be aware of that limitation as well.
so be aware of that limitation as well.
it takes time for these curves to fill in, but right now, there seems to be some sort of national scale effort going on to back-fill REALLY old cases.
we saw that here. https://twitter.com/boriquagato/status/1281686147819683846?s=20
we saw that here. https://twitter.com/boriquagato/status/1281686147819683846?s=20
everyone is trying to do open heart surgery with a steam shovel here.
the data is just awful and desperately misleading either though carelessness or deliberate skulduggery.
so question it.
dig. find these issues and make them clear.
that's how we get out of this hole.
the data is just awful and desperately misleading either though carelessness or deliberate skulduggery.
so question it.
dig. find these issues and make them clear.
that's how we get out of this hole.
