COVID19 Statistics

Passionate about UK politics and the global coronavirus response. 
Creating accessible charts to track the COVID-19 pandemic in the UK
Follow this Taaalk

4 followers

1,118 views

Joshua Summers
10:05, 08 Jun 20
So right now on Reddit and Twitter you are the COVID statistics person. Posting graphs every day that track the number of deaths in a range of settings, with this one being the latest:

xj8jwnagvh351 (1).png 996 KB

You have been tracking the death statistic more closely than most. What has been your experience of following it over the last few months? Has its progression gone as you've expected? Or have you been surprised by anything?
Thomas Day
19:51, 08 Jun 20
Hi Josh, yes it's something I started out just doing for myself as I became aware that was being reported as daily figures wasn't actually what happened in the past 24 hours. Having a bit more time than I would normally during lockdown I began creating these charts and found people were also finding them useful on Reddit so have continued. The useful feedback I've got has also helped to form them, they were pretty basic at the start! I've also seen it as a bit of a visual challenge to get as much information into one chart so it has lots of detail but is also easy to read.
Thomas Day
19:53, 08 Jun 20
In terms of what my experience has been following the data I've been most surprised by how "by the book" it has been. A steep rise followed by exponential decay - this means as the number declines the speed of the fall also declines. I'm working on a log chart at the moment which shows this more clearly. In this form the decline so far has been an almost perfectly straight line.
Thomas Day
19:57, 08 Jun 20 (edit: 19:59, 08 Jun 20)
This other thing I've found surprised by is how this has been so poorly reported not just by the government but also the media. I think there is such a rush to get the latest breaking information on how many deaths have happened today there has been little interest in reporting the accurate picture which you need time to establish. The chart on Sunday showed that the rolling average of deaths in hospital fell below 100 on the 30th May, this is a major milestone however as we are only able to establish this with some certainly 8 days out it doesn't get reported. Just a few days ago we were reporting that deaths were 300+, which given the trends in the chart is very unlikely.
Joshua Summers
10:15, 09 Jun 20 (edit: 10:36, 09 Jun 20)
I feel like your last point speaks to something present in the British media, and maybe Britain as a whole - we have an inability to sit with subtlety.
It seems the media likes to think in extremes, and that is difficult as the world is always more grey and muddled. When it comes to the reporting of deaths it does not surprise me that we want the the here and now instead of the true picture which may take more time to establish.
I agree about the mathematics, when I saw the first graphs coming out of Wuhan I was extremely surprised about the perfectly exponential nature of the infection curve (the numbers were slightly on the made up side of things though). (The blue line on this graph:)

PastedGraphic-3.png 351 KB

Behind the numbers you've been tracking lies an infection base. If you were to take clues from the death figures to build a story on how the infection base has risen and fallen in the UK, what do you feel has happened/is happening?
Thomas Day
16:38, 09 Jun 20
Yes the thirst to have the latest news and also social media has had an effect on the accuracy on how this has been reported, reporting what has happened 7+ days ago just isn't of interest, we want to know what happened today.
Thomas Day
16:40, 09 Jun 20
Also it's an interesting point you raise about the tendency of the media to push for extreme narratives, often in pursuit of clicks, rather than presenting a clear picture of what's happening which may not be as interesting.
Thomas Day
16:46, 09 Jun 20
I must admit I haven't done much around international comparisons and I do believe it is very hard to make them at the moment. There has been a lot of debate around this especially when the comparison graphs on the government daily briefings were dropped as we were leading. I do believe it is the case that these had been used as at first it showed us lower than other countries, as soon as we started to take the lead they were dropped. Really they shouldn't had been there from the start. Just looking at the numbers in Spain they dropped last week to 0 which got a lot of coverage, whereas what had happened is Spain had changed their method of reporting. It will take time to carefully, hopefully independently, analyse excess death figures from other countries until we have a clear picture of what has happened globally. 
Thomas Day
16:53, 09 Jun 20
Regarding the infection base you may wish to have a look at https://coronavirus.data.gov.uk/ . This has been something which has quite surprised me. The main reason I began to track the number of deaths is as I saw these are a hard figure which was best to follow the progress of the virus. Infections were based a lot on our ability to test and this was very low for quite a big chunk of the pandemic - there were lots of people who had this who weren't tested and so we didn't know about.
Thomas Day
16:59, 09 Jun 20
So it's quite surprising that when you look at the daily number of lab confirmed cases in England how closely it follows the deaths by actual day chart. The peak is only one day off the deaths chart for example and the sharp increase followed by exponential decay in the rolling average are remarkably similar. Personally I would have thought due to the lack of testing earlier in the pandemic numbers would be under recorded and cases while going down over time from the peak wouldn't fall in the same way. As testing capacity increased you'd expect to be picking up more cases even though overall number in the country would be falling. I must admit this isn't something I can explain.
Joshua Summers
14:33, 10 Jun 20
You might know better than me but I believe there has been a generosity used in the numbers when it comes to results. As far as I understand, the number of tests != the number of people tested. A swab to the throat and the nose of the same person counts as two tests not one, but I believe the reporting in terms of COVID-positive results is done on a per person basis, which would explain how even though the "number of tests" has increased, the number of positive cases/the ratio has fallen so much.
A few clues can be seen on the gov.uk website:
As of 9am on 9 June, there have been 5,870,506 tests, with 102,930 tests on 8 June.

289,140 people have tested positive.
You can see how they only insert the word "people" on the second line regarding positive tests. Then they have the following graph:

image.png 12.3 KB

And again, you can see how the difference between "Tests" vs "People tested" is present (and not resolved!). Again a fairly generous amount of ambiguity is used in the heading of the "Positive" chart, as it's not obvious if that's positive tests or positive people tested.
There has been a lot of discussion about a second wave, but from what you're saying about the steady fall of deaths in the UK, it feels like it certainly has not manifested in these statistics. Do you feel like the second wave conversation is overblown, or - perhaps as deaths are the most lagging indicator - do you feel like the risk level could be high?
Thomas Day
16:45, 11 Jun 20
Yes it's hard to fully interpret the testing data due to some of the some of the discrepancies you have highlighted. This is why I was surprised that the chart on the government website of positive results follows the deaths by day chart so closely.
Thomas Day
16:49, 11 Jun 20
I also think were the tests are collected are an important indicator. For a large period of the pandemic tests were just being conducted in hospitals with those who were admitted, clearly you'd expect this to trigger more positive results than when you are testing the wider community. Overall I think tests, while very important, are not a good measure you track the spread of the pandemic, until you have certainly that a majority of the population are being regularly tested.
Thomas Day
17:09, 11 Jun 20
Regarding a second wave, this is something which has been mooted due to what has happened in previous pandemics and certainly in some states of America at the moment there doesn't seem to be a slowing that has been experienced in many European countries and a second wave is feared.
It is important to highlight I'm no epidemiologist however I do keep a keen watch on what's happening with the coronavirus so this is purely my opinion.
Clearly opening things up again you would expect transmission to rise, and therefore a second wave. There are a number of factors which are in play here however. The warmer temperatures in Summer do help slow this virus, people even though things are opening up are being more careful and a lot of social distancing measures are now in place, a proportion of people now have antibodies to the virus (supposed to be 20%+) in London which removes them from spreading, and there are now a number of different strains of coronavirus in transmission. All viruses survive by reproducing and often there is a slight mistake in the genetic code, this is what leads to mutations. 
Over time as a virus spreads it will be the mutations which are able to spread the most which are the most successful and this tends to favour the weaker strains. Any stronger mutations which remove the host from society or even worse kill the host are not going to have the same ability to spread. A mutation of coronavirus where most people are asymptomatic will have the most success of spreading. 
It may be optimistic but I am hopeful lockdown has helped protect us from the stronger strains which were around in the early part of the year and now coming out of lockdown, with all the additional factors considered, it will be weaker strains which help keep deaths from the virus low and ultimately kill off the virus with or without a vaccine.
Again I will reiterate this is purely my opinion based of numerous academic sources online and is not the expert view of an epidemiologist which you really need here, so take with a pinch of salt.
Thomas Day
17:27, 11 Jun 20
I think in terms of the most timely measures to identify a second peak will first be hospital admissions followed by deaths. If the virus weakens and there is a rise in asymptomatic cases it may well be the case that the amount testing positive to the virus do increase even though this won't register on hospital admissions or deaths - however again it will depend how testing is carried out. At the moment you can get a test if you have symptoms, if you don't then these tests aren't being carried out. I would hope with increased testing capacity there can be testing on a large scale of specific areas and settings allowing us get a more accurate picture of what's going on.
Joshua Summers
11:07, 12 Jun 20
Yes I have found testing to be a rather frustrating thing to observe... I understand that we have been ambitious but I think we should think in factors more than that - 10 times as ambitious at least. I want to live in a world where COVID testing trucks are driving down every street on a recurring basis, and getting people out of their homes one by one to perform tests.
The news today is all about Conservative back benchers asking for the 2m rule to be changed to a 1m rule, but ultimately the public confidence has to be there to open up the country in any meaningful way. I don't think it is right now (maybe a decent period without a rise in cases will make that happen). A testing strategy which was aggressive and present would change that.
I'd say that this microcosm of the unambitious (and therefore more affordable) approach to testing is a systematic problem with the government's approach. We do not go extreme enough fast enough, and paradoxically with this virus that is exactly the approach that is needed to squash it. Cost saving (whether that be allowing a few more weeks of economic activity before lock down, or not going to the moon with tests) ends up costing more.
Joshua Summers
11:10, 12 Jun 20 (edit: 14:37, 12 Jun 20)
I do think the UK is in a bit of a muddle right now. The COVID cat is out of the bag, and we missed our chance to keep it in. I think at the start of all this the most obvious mistake was made (not shutting borders + locking down early). I believe it's much harder to know the obvious thing to do today.
If you found yourself in Number 10 with all the power, what would you be doing or thinking about right now?
Follow this Taaalk

4 followers

1,118 views

Start your own Taaalk, leave your details or meet someone to Taaalk with.