Daily numbers, graphs, analysis thread 2

[Barracker]

A follow on thread from here

Please try to keep it data driven, factual and civil.

www.worldometers.info/coronavirus/country/uk/

If this is the case, then what we are observing each day is not even close to being current, and we may have to wait days before we properly understand what the true toll is and how it is evolving.

This is one of my concerns. I dread to think what the current figures might actually be.

As per my post first thing this morning, none of my trust's deaths have been included in national totals yet due to some confusion about where these were supposed to be reported. They will now (or already have been - I've been off) be included in one go. And I don't suppose for one minute that this situation applies to just one trust.

Thank you for that BBC news article posted at 08.57 today Baaaahhhh. Very clear.

"new cases going up by 5% which is more positive."

And to interpret that as a trend would you need to know how much the number of tests carried out had also increased?

@BigChocFrenzy I think this

docs.google.com/spreadsheets/d/1eTKeK9vRxgw0KhvKxPCaDrfaHnxQP-n9TsLzsEymviY/htmlview#

shows what you want .

As an aside I can't work out where @Murtaman who compiles this gets the testing capacity figure from or why there always seems to spare capacity .

U.K. testing numbers

Testing capacity for inpatient care in England currently stands at 12,799 tests per day.

I wonder what the capacity is for other tests, such as frontline staff etc.

As per my post first thing this morning, none of my trust's deaths have been included in national totals yet due to some confusion about where these were supposed to be reported. They will now (or already have been - I've been off) be included in one go. And I don't suppose for one minute that this situation applies to just one trust.

Holy shit!

Like, completely understandable and no blame intended (frontline work comes first!) but this could mean we’re in for some shocking jumps.

It's so misleading though because it also says only 7551 people were tested, so many of those must have been repeats. The percentage of positive tests today is really high if you take into account the number of people tested rather than overall test numbers. 55-60% ish?
So gutted about these increasing numbers.

If 10,000 tests are carried out, that does not mean that 10,000 people have been tested.

I wonder if they are going to say how many individuals have been tested?

Real time results have never been an NHS strong point. This much I can attest to through work.

I also think there’s possibly a delay in results so people are dying before a test confirms their covid-19 status so someone needs to marry up the figures retrospectively.

It sounds simple enough but my experience tells me that there will be teething problems and miscommunications/different priorities at multiple levels.

Yes people will be dying before swab results are back. They're taking quite a while.

In my trust they're using x-ray/ct has a diagnostic guide as the appearance of Covid19 is very unique.
Then they confirm with swabs.
Many admissions to itu/wards are on diagnostic imaging alone at the moment and swabs are concrete recognised data.

@chomalungma
If 10,000 tests are carried out, that does not mean that 10,000 people have been tested.

some people could have been tested 2 or more times, so no?

One question.

If there's a lag and some of these mortalities were days ago. Would that not possibly show our death rate increased earlier than we thought and therefore we are going to reach the peak sooner than figures show?

I'm just still so floored by the number of deaths over the past 24 hours that have been reported. It's jumped a lot in the past week and we know there's worse to come

It shows the death toll increased sooner and higher than thought, yes.

But it doesn't follow that we will reach peak sooner.
The peak isn't dependent upon us reaching a certain number and then declining.
The peak is a function of how long ago we locked down, and to what degree the lockdown is effective.
It's more likely that the peak will be stretched out further away, and higher, since the numbers are on borrowed time.

Using other countries as templates, lockdown yields results after maybe 3 weeks. A lot now depends upon whether their figures were promptly reported, or had severe lags, like us.

itsgettingweird that's sort of what I was asking earlier.

Is it possible to access any data on when each death actually occurred in order to chart a graph that might show a more accurate increase in numbers?
In my head, and I'm struggling to articulate this, but if there have been 684 deaths reported today but some of them occurred on previous days - our death total is still the same but the curve of the graph wouldn't be as steep with such shocking jumps in numbers? Which wouldn't make it look as bad as it does now?

Thank you, @peridito

Moo but there will be some deaths that occurred yesterday/today that won't be counted for another week - pos 2 weeks.
iyswim?

Moochops in that scenario, it's possible that 684 deaths are reported today, the majority of them are swept up from ages ago, however, the majority of yesterday's deaths are not included, and they in their turn will be swept into future figures.
The problem is, for as long as we are on the upward growth phase of a exponential curve, the numbers missing from yesterday are likely to be a great deal higher than the older swept-in data that was included.

It would be so important right now to have all the data date stamped. It's obvious that reporters are getting access to this date stamped data, but noone has yet been able to locate it for public scrutiny.
There has only been the one example at the beginning of this thread that was in the papers a few days ago.

I have no idea WHERE this data can be accessed, although obviously it can be, for the Guardian to be able to state that only 84/561 deaths were actually from the day they were reported to be. That's 15%.
85% of that days report was older data.
Those proportions worry me.

If we had datestamped data it could be used to map actual deaths by actual date, and we'd see the true graph.

Agree, Barracker

Thanks for these threads, reading with interest (mathematician who has been astonished at the lack of understanding in general public, politicians and journalists of basic calculus i.e. which derivative of a function/sequence are you talking about at any given time). Logged on for first time in ages to give same bleak comeback to Moochops as you, Barracker!

Thanks barraker that makes sense.

Just read that Kate Garraway DH is in ICU and seriously ill from CV she expressed (maybe last week or end of week before?) concern she'd met Prince Charles a few days before he'd been confirmed. She's also showing mild symptoms.

It's so sad but I'm hoping the more people who are in the public eye are reported the more people will take this seriously and stay in this weekend rather than enjoy the sunshine on questionable loopholes and walks. People need to actually realise just how much this virus doesn't discriminate. The best private medical care is not going to help anyone.

Over the last week the growth rate in the UK is noticeably steeper than in Italy at a comparable point on the curve.

That's a terrible graph though / look at the left hand axis!! You need to have the intervals equidistant to be able to read if properly.