Covid 19 Part XXI-27,908 in ROI (1,777 deaths) 6,647 in NI (559 deaths)(22/08)Read OP

ACitizenErased

gmisk wrote: »

I have a masters in computer science, and worked in analytics for about 4 years....so yes I know what an outlier is..no need to be so condescending....you don't have to totally discount it, they are still confirmed cases and not an outlier given the entire data set.

I'm talking statistics - an outlier exists within the dataset which is skewing the mean. Do you have an alternative view on this?

drunkmonkey

is_that_so wrote: »

I am curious who he is. If he's that crackpot Swede who said we'd have lots more cases I'd be reaching for the salt. As for pubs remember the Tyrol resort and the skiing in February, with the packed bars, the Korean nightclubs and how many cases came out of them? That too is evidence.

It was session 3 I watched, i'm not sure if you can watch a repeat somewhere, trust me this guy isn't swedish and an idiot like your trying to imply to support your assumptions. His name illudes me and the full list of attendees wasn't published, maybe it'll be online later I'll keep an eye out so you can listen to experts yourself.
https://www.oireachtas.ie/en/committees/schedule/

Blondini

ACitizenErased wrote: »

Because there's an outlier in the data...... I explained this above.
An example:
Naive interpretation of statistics derived from data sets that include outliers may be misleading. For example, if one is calculating the average temperature of 10 objects in a room, and nine of them are between 20 and 25 degrees Celsius, but an oven is at 175 °C, the median of the data will be between 20 and 25 °C but the mean temperature will be between 35.5 and 40 °C. In this case, the median better reflects the temperature of a randomly sampled object (but not the temperature in the room) than the mean;

I love your work here ACE but all outliers cannot automatically be dismissed.

ACitizenErased

Blondini wrote: »

I love your work here ACE but all outliers cannot automatically be dismissed.

I didn't say dismiss it? I said it's skewing the mean. It obviously cannot be dismissed. I never once said dismiss it.

froog

froog

ACitizenErased wrote: »

Because there's an outlier in the data...... I explained this above.
An example:
Naive interpretation of statistics derived from data sets that include outliers may be misleading. For example, if one is calculating the average temperature of 10 objects in a room, and nine of them are between 20 and 25 degrees Celsius, but an oven is at 175 °C, the median of the data will be between 20 and 25 °C but the mean temperature will be between 35.5 and 40 °C. In this case, the median better reflects the temperature of a randomly sampled object (but not the temperature in the room) than the mean;

i'll help you out here. that line is trending up.

ACitizenErased

Clearly cannot make a point on here without words being twisted and taken out of context, batten down the hatches, I'm out.

stephenjmcd

froog wrote: »

Are you actually going to add to the conversation / debate or just keep posting the likes of the above, referring people to the trump administration & jumping up and down to go back in to phase 1 and 2.

Christ I'm glad I gave up doing any statistical analysis on this forum when this is what your up against.

People make mistakes, stats change daily as we're seeing

drunkmonkey

s1ippy wrote: »

I'm sorry now like but you're going to have to give a name for this utter lunatic if he exists at all.

Professor Carl Heneghan EBM & Director of CEBM at the University of Oxford.

Sure what with an utter lunatic like that know:rolleyes:

Renjit

Guys, we still good this week. See ya next week.

The Unbearables

froog wrote: »

Sums up those in deep denial very nicely.

froog

stephenjmcd wrote: »

Are you actually going to add to the conversation / debate or just keep posting the likes of the above, referring people to the trump administration & jumping up and down to go back in to phase 1 and 2.

Christ I'm glad I gave up doing any statistical analysis on this forum when this is what your up against

i actually took the time to graph the 7 day moving average and helpfully analysed it to show people it is trending up, whilst pointing out the absurdity of picking a tiny section of that trend to show it is trending down.

what have you done?

drunkmonkey

is_that_so wrote: »

I am curious who he is. If he's that crackpot Swede who said we'd have lots more cases I'd be reaching for the salt. As for pubs remember the Tyrol resort and the skiing in February, with the packed bars, the Korean nightclubs and how many cases came out of them? That too is evidence.

Watch this and let us know what you think...

https://www.youtube.com/watch?v=H0CQ4nd7gD4

MOR316

I'm amazed some of you have managed to leave the house at all in your lives incase the rain melted you.

stephenjmcd

froog wrote: »

i actually took the time to graph the 7 day moving average and helpfully analysed it to show people it is trending up, whilst pointing out the absurdity of picking a tiny section of that trend to show it is trending down.

what have you done?

plenty before you came onto this thread.

Gave up doing any analysis and posting it due to posters in a similar vain to yourself, it wasn't worth the hassle you get for it.

Numbers are going to go up and down, daily trends, weekly trends all going to fluctuate. But sure go back a few phases

Good luck

Non solum non ambulabit

froog wrote: »

i'll help you out here. that line is trending up.

Would you mind posting that graph on Saturday after the numbers are released? Be interested in your commentary then.

darjeeling

BringBackMick wrote: »

Is there any evidence or scientific discussion with regards to anectodal evidence that the 'second wave' of Covid19 in Europe is not resulting in a big uptick in hospitalisation ?

We're seeing the beginnings in Spain, which has the biggest epidemic in Western Europe right now. Hospitalisation admissions due to covid were at ~100 per week in June compared with over 800 per week as of yesterday. Weekly ICU admissions are up to around 50 per week from under 10 in June.

But the case numbers across Europe - even though they are definitely increasing - are still likely small relative to the March / April peak numbers because testing was missing 90-99% of cases at the peak, compared to more like 50% now.

spookwoman

I thought Wiki articles were not allowed in here

froog

stephenjmcd wrote: »

plenty before you came onto this thread.

Gave up doing any analysis and posting it due to posters in a similar vain to yourself, it wasn't worth the hassle you get for it.

Good luck

i don't know what threads you've been reading tbh, all i've seen is genuinely worried people getting mocked relentlessly by people who miss their pint.

bettyoleary

MOR316 wrote: »

I'm amazed some of you have managed to leave the house at all in your lives incase the rain melted you.

OOOh I bet your kids are doing your head in right now.

Non solum non ambulabit

stephenjmcd wrote: »

plenty before you came onto this thread.

Gave up doing any analysis and posting it due to posters in a similar vain to yourself, it wasn't worth the hassle you get for it.

Good luck

Your data is missed.

ACE is holding back the hoards of panic merchants daily with his statistics provided with context.

Consistently miss interpreted by some, intentionally I would contend.

stephenjmcd

Non solum non ambulabit wrote: »

Your data is missed.

ACE is holding back the hoards of panic merchants daily with his statistics provided with context.

Consistently missed interpreted, intentionally I would contend.

Thanks Non,

Was more hassle than it was worth. Pop in and out to see what's new in here. Still keep track of stats myself but its more of a personal project now, hoping to do some proper analysis and plot the data when this is all eventually passed us.

Will share it out then for sure.

Jackman25

froog wrote: »

i don't know what threads you've been reading tbh, all i've seen is genuinely worried people getting mocked relentlessly by people who miss their pint.

Clown.

JP Liz V1

It's time I asked so wtf is an outlier/outliers

mr zulu

spookwoman wrote: »

People are surprised in the increase in numbers? oh please where does it show people are surprised in the numbers no where but anyone who says 92 is not good are doom mongers.
Just because a some health official says oh the numbers will increase, I keep seeing the same meme in my head .

froog wrote: »

the rising cases.

froog wrote: »

but it's going down in this bit here :pac:

froog wrote: »

i don't know what threads you've been reading tbh, all i've seen is genuinely worried people getting mocked relentlessly by people who miss their pint.

Non solum non ambulabit wrote: »

Your data is missed.

ACE is holding back the hoards of panic merchants daily with his statistics provided with context.

Consistently missed interpreted, intentionally I would contend.

Yes your right,when they smell bad news,they come out in a pack,keep up the good work,ACE and Steve,I only found the ignore option lately,I’m putting it to good use.

KrustyUCC

stephenjmcd wrote: »

plenty before you came onto this thread.

Gave up doing any analysis and posting it due to posters in a similar vain to yourself, it wasn't worth the hassle you get for it.

Numbers are going to go up and down, daily trends, weekly trends all going to fluctuate. But sure go back a few phases

Good luck

I can only imagine the very large societal and economical shocks to the country going back to phase 2 not to mind phase 1

The numbers are not there to warrant this countrywide

Boggles

JP Liz V1 wrote: »

It's time I asked so wtf is an outlier/outliers

a younger rock formation isolated among older rocks

I think they are talking about rocks.

I'm not sure though.

JP Liz V1

ACitizenErased wrote: »

Clearly cannot make a point on here without words being twisted and taken out of context, batten down the hatches, I'm out.

Take Care Come Back ACE, your data appreciated

WhiteMemento9

KrustyUCC wrote: »

I can only imagine the very large societal and economical shocks to the country going back to phase 2 not to mind phase 1

The numbers are not there to warrant this countrywide

I am not saying we should do one thing in particular as we have experts better placed to make these decisions.

Your line of thinking though is not very good. You make these decisions early if nesscary to curb and cut things off at source so you only need to revert for short periods rather than letting the virus get out of control as happens with exponentials and then having to endure very long sustained periods of lockdown which do far, far more damage.

ACitizenErased

JP Liz V1 wrote: »

It's time I asked so wtf is an outlier/outliers

An outlier is a piece of data (ie one day out of 7) that causes a skew in the data. In this case the 174 cases on one day dramatically drives the 7 day average upwards until that day is removed from the 7 days eventually.

Jim_Hodge

Non solum non ambulabit wrote: »

Your data is missed.

ACE is holding back the hoards of panic merchants daily with his statistics provided with context.

Consistently missed interpreted, intentionally I would contend.

Ah hold on. They are not his statistics for a start. I do value the daily update of these stats though. However, you cannot skew the stats by removing the counties driving the national figures. Next week it could be different counties. Do we exclude them too? The 174 cases will drop off, yes, and then the 7 day average will drop significantly if we hold to current numbers. But this virus will see jumps from time to time and we can't exclude them all.

I've studied statistics at professional level so can appreciate the arguments but don't accept the exclusion of those that don't fit the narrative.

JP Liz V1

Boggles wrote: »

I think they are talking about rocks.

I'm not sure though.

ACitizenErased wrote: »

An outlier is a piece of data (ie one day out of 7) that causes a skew in the data. In this case the 174 cases on one day dramatically drives the 7 day average upwards until that day is removed from the 7 days eventually.

I thought it was a The Walking Dead reference I didn't get

KrustyUCC

WhiteMemento9 wrote: »

I am not saying we should do one thing in particular as we have experts better placed to make these decisions.

Your line of thinking though is not very good. You make these decisions early if nesscary to curb and cut things off at source so you only need to revert for short periods rather than letting the virus get out of control as happens with exponentials and then having to endure very long sustained periods of lockdown which do far, far more damage.

That's why we have localised lockdowns when warranted like in LOK but not a country wide one

s1ippy

https://twitter.com/oconnellhugh/status/1293497263868776448?s=20

spookwoman

Jim_Hodge wrote: »

Ah hold on. They are not his statistics for a start. I do value the daily update of these stats though. However, you cannot skew the stats by removing the counties driving the national figures. Next week it could be different counties. Do we exclude them too? The 174 cases will drop off, yes, and then the 7 day average will drop significantly if we hold to current numbers. But this virus will see jumps from time to time and we can't exclude them all.

I've studied statistics at professional level so can appreciate the arguments but don't accept the exclusion of those that don't fit the narrative.

We might not always agree but you are 100% correct you cannot exclude data because as you said it doesn't fit the narrative.

Jim_Hodge

ACitizenErased wrote: »

An outlier is a piece of data (ie one day out of 7) that causes a skew in the data. In this case the 174 cases on one day dramatically drives the 7 day average upwards until that day is removed from the 7 days eventually.

But who decides on what constitutes an outlier? Will you drop today's 92 early next week, as it will be the outlier in the trend?

JP Liz V1

s1ippy wrote: »

https://twitter.com/oconnellhugh/status/1293497263868776448?s=20

Is Hugh, a stalker pap for Harris

Non solum non ambulabit

Jim_Hodge wrote: »

Ah hold on. They are not his statistics for a start. I do value the daily update of these stats though. However, you cannot skew the stats by removing the counties driving the national figures. Next week it could be different counties. Do we exclude them too? The 174 cases will drop off, yes, and then the 7 day average will drop significantly if we hold to current numbers. But this virus will see jumps from time to time and we can't exclude them all.

I've studied statistics at professional level so can appreciate the arguments but don't accept the exclusion of those that don't fit the narrative.

He has been providing greats stats and analysis for months and you are going to hang him over a few posts tonight when he was clearly put out by the constant picking at his data?

I will defend him as he is the most valuable contributer to this thread. With out him I expect many would leave.

s1ippy

He just quickly copies the numbers from the official website??

Goldengirl

MD1990 wrote: »

I had similar.
Took me about 4 weeks to get my energy.

Had no taste & smell & couldnt do anything for a week.
My taste & smell may not have fuly returned yet.

4 of us had a godawful dose from last week of January to mid March , one after the other .
Myself and Oh were sick for 3 weeks with it but didn't meet the criteria fir testing , and then zi was negative when tested as a healthcare worker , 2 weeks after symptoms gone ...worse than useless, but around the time of the delayed testing .
All working in essential areas since and not a sniffle , fingers crossed .
But if it was it , I found it very scary , and a lot worse than any flu I have ever had .

spookwoman

s1ippy wrote: »

He just quickly copies the numbers from the official website??

and from other people.

lee_baby_simms

JP Liz V1 wrote: »

Is Hugh, a stalker pap for Harris

Blue Hugh. A pillar of objective journalism.

Boggles

s1ippy wrote: »

https://twitter.com/oconnellhugh/status/1293497263868776448?s=20

Stopped on a double yellow line and wearing his mask on his chin.

Boggles

Ohmeha wrote: »

7 day rolling average is currently up marginally as of today, however as a health warning the case numbers are yo-yoing up and down this week and by Saturday 15th when you can exclude the large numbers cases on Fri 7th and Sat 8th then the 7 day average will trend downwards if the cases for the next 2 days average at even 80 per day, we'll see anyway

Day Month Date Cases 7 Day
Friday July 24th 20 16.71
Saturday July 25th 24 17.14
Sunday July 26th 12 17.43
Monday July 27th 11 18.14
Tuesday July 28th 40 18.71
Wednesday July 29th 14 18.29
Thursday July 30th 85 29.43
Friday July 31st 38 32.00
Saturday August 1st 45 35.00
Sunday August 2nd 53 40.86
Monday August 3rd 46 45.86
Tuesday August 4th 45 46.57
Wednesday August 5th 50 51.71
Thursday August 6th 69 49.43
Friday August 7th 98 58.00
Saturday August 8th 174 76.43
Sunday August 9th 68 78.57
Monday August 10th 57 80.14
Tuesday August 11th 35 78.71
Wednesday August 12th 40 77.29
Thursday August 13th 92 80.57

TBF the only unique stats that I see being posted is from this lady / gentleman.

And thanks a million for doing it.

Jim_Hodge

Non solum non ambulabit wrote: »

He has been providing greats stats and analysis for months and you are going to hang him over a few posts tonight when he was clearly put out by the constant picking at his data?

I will defend him as he is the most valuable contributer to this thread. With out him I expect many would leave.

Would you give over. I'm hanging nobody and have complimented his posts regularly: However the stats are available elsewhere. Analysis is, and always be, subjective and today it is being manipulated, by excluding pertinent data, to give a biased outlook. I'm all for the stats and analysis and the picture will certainly change frequently but let's look at the entire picture.

I think we're grown up enough to be able to discuss the figures and any analysis pertaining to them.

spookwoman

Non solum non ambulabit wrote: »

He has been providing greats stats and analysis for months and you are going to hang him over a few posts tonight when he was clearly put out by the constant picking at his data?

I will defend him as he is the most valuable contributer to this thread. With out him I expect many would leave.

He is not an official spokesperson for the HSE and Government or an expert on stats, the virus etc.
He contributes to the thread but that does not mean what he posts is always right and people are allowed to question him on it.

JP Liz V1

Boggles wrote: »

Stopped on a double yellow line and wearing his mask on his chin.

He was probably asking for a cigarette

Speak Now

s1ippy wrote: »

https://twitter.com/oconnellhugh/status/1293497263868776448?s=20

Hope they got the breathalyser out

gmisk

JP Liz V1 wrote: »

Is Hugh, a stalker pap for Harris

It's almost as if his wife Kate oConnell is big in FG....

Goldengirl

Thierry12 wrote: »

Wouldn't read much into it

The tests we use are high specificity,the downside of that high specificity is that we miss many positives

Also depends on the day of infection.

The likelihood of getting a false negative test result is way higher on the 1st day of infection, than on the 7th day

False positive results ) are thought to be much less of an issue with pcr tests compared to false negative results.

It's a bit like law, we can only convict when 100% sure, 90% isn't enough to convict

The test itself is super sensitive.They have to be taken correctly, and even if they are taken correctly the virus has to be in the spot where the sample is taken.

All this 99% accurate is in perfect conditions, performed by a professional.

In the nursing homes for example the nurses that work there are doing the tests, they were shown for 5 mins how to perform the test, they are so inconsistent from nurse to nurse, they might as well not bother testing.

I've been tested 4 times now and each time was different, it's a hard test to give

I think if it was a nurse taking it I'd be very happy . They are professionals , who are mostly used to doing similar swabs from training onward.
It's the self testing that is a total disaster , ie in UK , and postal tests .