Data and analysis thread, started 12 November

[NoGoodPunsLeft]

Previous thread here:

Data and Analysis Thread, started Oct 29 www.mumsnet.com/Talk/coronavirus/4064113-Data-and-Analysis-Thread-started-Oct-29

Regular lurker but I frequent poster, didn't want to lose the threads.

And parents told they will be fined if they keep children at home

[quote peridito]@FeelingBIue

my son is a TA in an Academy run primary in Newham where cases continue to rise .

www.newham.gov.uk/downloads/file/2500/public-dashboard-18-11-20-js-pdf

Each year group of 3 classes = one bubble .Lunch time = sandwiches in classroom supervised by dinner ladies with class teacher in staff room .

A TA went home sick on a Tuesday because of uncontrollable coughing .
Confirmed positive on the Thursday . Is in hospital v unwell .

School still open ,no bubbles isolating at home .[/quote]
Thanks Perdito that link is really useful.

So although Newham +ve numbers are rising the number of tests carried out are rising at a faster rate. Also states borough has an increasing rate in the over 65s - would they say if schools are the driver-up of numbers though? No comfort to the poorly TA and all the sick patients in hospital but looks promising that the rate of infection in Newham shows falling as a percentage of tests carried out.

My borough gives no info on tests carried out if Newham can do it then figures must be available. Going to have a better search in the statistics.

What's the general opinion about this

imperialcollegelondon.github.io/covid19local/LTLA_public/Richmond_upon_Thames.html

?

For Richmond or the general page?

Haven't seen that before, interesting! But scary when I searched for our area.

Mine too August! They're not very optimistic about my town 😱 I wonder whether it's in the toolbox for deciding what happens and where after 2/12.

@TheSunIsStillShining

What's the general opinion about this

imperialcollegelondon.github.io/covid19local/LTLA_public/Richmond_upon_Thames.html

?

I have to confess I can't understand most of the graphs in that link TheSunisStillShining.

That said the R-rate one for my borough suggests it is very slowly falling even though cases are rising. Does that mean it's likely increase in testing driving the increased +ve numbers?

I don't understand what's the point of graphing projections! Unless you're studying accuracy of forecasts which is a separate thing.

I don't know what to believe/consider to be valid or even vaguely informative.
On one hand Richmond mobility data* is genuinely stupid. In London mobility is down 22& and yet in R. people act as if nothing is wrong, only 6% drop in mobility. And couple this with the imperial college's dashboard of R falling - makes no sense.
And then I look at actual numbers (eg. gov) and the local msoa number are on the rise...

*www.datarich.info/covid-19/#COVID19local
you can see other London boroughs as well

Interesting graphs @thesun But I'm not sure it's really a valid thing to do to try to make projections for the future in the basis on past trends when we know that the things that cause infection rates to rise and fall have changed significantly in the last few weeks (causing the data they're using to project from) and are very likely to change in the future too. This is particularly clear when you look at areas where R is close to 1 and they're charts show that absolutely anything could happen! I think if they'd made projections that way a couple of months back they'd have entirely failed to project what actually followed in different areas. You need a more holistic, intelligent approach to understanding what's happening in local areas.

@Feelingblue It's quite possible for the R rate to fall while cases are rising. If the R rate falls from, say, 1.8 to 1.1 cases will go on rising all they time (because even when R=1.1 each person is on average infecting 1.1 others) They'll just be rising more slowly as R falls closer to 1. (This confused a lot of people in the summer when R rose from about 0.7 to 0.9 but cases went on falling, which of course they would until R got to 1 exactly)

You can’t just look at the whole North East or South West. So true. I live in the South West, last night the local news had a piece about rates in Torbay which a week or so ago were the highest in the South West I think. The numbers are going down but they said the problem is Brixham, neighbouring Paignton and Torquay are about half the rate of Brixham. There is hardly a gap between Brixham and Paignton but the figures were something like high 300s per 100k in Brixham and high 100s in Paignton. If it is that big a difference in Torbay there isn't much of an indication of what is happening in the South West in general.

I find it interesting how much the ages of cases is varying by area, it shows there are very different patterns.
E.g. Bolton it's almost every age group right up to 50s, Newham almost exclusively 25-35yo: victimofmaths.shinyapps.io/COVID_Cases_By_Age/.

County Durham's an interesting graph.
Dropping, except on the 90% CI there's a possibility of it rising quickly. All the other areas I've looked at are fairly consistent.

Torbay meanwhile, much older set; 40-55 year olds leading the infections.

@TheSunIsStillShining - it’s interesting when you compare Richmond to the surrounding boroughs which have much more depressing predictions ( I’m really hoping it doesn’t prove accurate for mine).

I think Torbay has an older demographic generally, although I suppose alot of that is older than 55. I wonder what is going on in Brixham with 40 to 55 year olds?

I wonder if the fishing industry is a factor? Food processing facilities have been involved in other places. I know there is a fishing fleet based in Brixham plus the fish market. Don't know if there are any fish processing places.

How on earth have Germany only reported 4K cases today??? They were 23k + yesterday. They don’t report by state/province do they?

Yes, their reported numbers increase through the day. This won’t be the final total.

19875 / 341 compared to 26860 / 462 last Saturday.

Definitely looking like we're turning a corner now.

@littlestpogo - howdy neighbour! I wouldn't pay it too much attention, the predictions are so wildly large in scale as to be virtually meaningless.

Thanks for the explanation on the R-rate Firefliess, at least we seem to be infecting each other at a lower rate, better than nothing.

@Mrex - yes I thought that re the predictions

@TheCountessofFitzdotterel - thank you. I’ve never noticed that before and I’ve been checking that site almost daily since March!

@littlestpogo - howdy neighbour! I wouldn't pay it too much attention, the predictions are so wildly large in scale as to be virtually meaningless

I think entirely meaningless for London boroughs, suggesting that Richmond has R confidently below 1 and Kingston has it confidently above 1 is just gibberish, there's no way the model can be sensitive to that, it doesn't have the information. The model might work better for Cornwall say, but otherwise there simply isn't the information to discriminate where the cases are.