|
Smirr posted:I don't know if the graphs are intentionally misleading (I suspect so), but they are misleading. Anyway, I guess this did turn into a meltdown after all vv Think they are just trying to make it more readable (i.e. less distance between curves than in linear scale) without using slightly awkward normalised values. Remember that a lot of people are not great at reading graphs like that in the first place.
|
# ? Apr 14, 2020 22:21 |
|
|
# ? Jun 8, 2024 07:41 |
|
elbkaida posted:Think they are just trying to make it more readable without using slightly awkward normalised values. Remember that a lot of people are not great at reading graphs like that in the first place. I don't mind the x-axis. Starting from case number whatever is completely fine to me. My beef is only with the y-axis. e: oh, you edited. If they want readability, which really is a concern, they could do a log plot of total number of cases. That would still show a downwards trend for countries that deviate downwards from exponential growth without implying that the top of the curve has been reached Smirr fucked around with this message at 22:25 on Apr 14, 2020 |
# ? Apr 14, 2020 22:23 |
|
Mithaldu posted:y'all should draw example pictures I adjusted the FT curve for a more realistic assessment of the situation.
|
# ? Apr 14, 2020 22:28 |
|
because of all this I make it a goal every day to look at 0 corona related graphs
|
# ? Apr 14, 2020 22:28 |
|
oliwan posted:because of all this I make it a goal every day to look at 0 corona related graphs But its a mix of "How to lie with statistics" and "How to fail to present data". Anyone know whether the RKI provide a CSV with the data they use for the dashboard, and where to find it?
|
# ? Apr 14, 2020 22:32 |
|
If you want to show a graph of new infections, a log scale is the only reasonable way, because in an uninhibited growth, this number would keep doubling and the early numbers would be unreadable. It doesn't matter if this goes over to the second or third derivative, since the curvature of these are pretty much identical in an exponential growth. As soon as any derivate starts to get negative, we can at least see that it stopped growing exponentially, which is the important mark. A lot graph turns this into passing a line. But I am speaking as a mathematician. Strictly speaking, since the total population is limited, with the right factors, a polynomial growth would be worse than an exponential one. Obviously, on that scale (1+10^(-10000))^x < x^(10000). That first term means 1,00000...0001 with 9999 zeros in between the ones. cant cook creole bream fucked around with this message at 22:56 on Apr 14, 2020 |
# ? Apr 14, 2020 22:39 |
|
Zwille posted:Have ya tried regular milk before they switched to Länger Haltbar? Yeah, I don't see the difference but I might be broken.
|
# ? Apr 14, 2020 23:23 |
Ika posted:But its a mix of "How to lie with statistics" and "How to fail to present data".
|
|
# ? Apr 14, 2020 23:32 |
|
DTurtle posted:https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv That's the john Hopkins data, not the RKI data (Which accounts for the lag between tests being done and reports arriving).
|
# ? Apr 14, 2020 23:46 |
Ika posted:That's the john Hopkins data, not the RKI data (Which accounts for the lag between tests being done and reports arriving). Here you go: https://www.arcgis.com/home/item.html?id=f10774f1c63e40168479a1feb6c7ca74
|
|
# ? Apr 15, 2020 01:34 |
|
Mithaldu posted:I remember something about a super spreader event in austria where everyone then basically spread over europe? Maybe just most cases that happened there exported themselves to elsewhere. No, the Norwegian server in Kitzloch has turned out to be just a sensational news story. He was infected, but he wasn't even the first case, just the first to go to the doctor from the bar as case tracing has shown. Playing beer pong with your mouth turns out to be a bad game when you have a virus pandemic starting up, and this was a favourite game in the Kitzloch. Combine this with being infectious 2-3 days before you show any symptoms or feel ill and you have a perfect virus spreading opportunity. Current tracing puts some Swiss and Italian tourists as the likely infection vector (and Tyrol aggressively advertised for Italian tourists to go to Tyrol after the South Tyrolian Schigebiete closed down because of Corona), and such funny games in tightly packed bars and even more packed Schilift Gondels spread it like mad. The whole way the Tyrolian authorities and Schilift-Barone and Hoteliers dealt with the situation was just disgraceful, callous and killed people and I hope they get sued into oblivion for it.
|
# ? Apr 15, 2020 05:35 |
|
This is Europe, not the US, when has anybody ever been sued and properly punished for something that wasn't a criminal activity? Edit: Pack it in folks, pandemic is over. TP in massive stocks in the supermarket again and ...excuse me I need to squeeze past here instead of going around you... distancing doesn't seem to matter any more. The people have decided, this poo poo is over. Seriously I just went shopping and it was a shitshow. People did not keep their distance despite the markers on the floor and the signs, plus enough space as the place was practically empty. Total of 3 masks (1 mine) between about 30 people inside, no masks on anyone outside at all. That was different last week. I guess all this talk about "planning to open up" had the unpredictable effect of nobody giving a poo poo anymore, who would've thought?I (Un)fortunately we never reached the point were everybody knew somebody who was severely ill or died, which would have driven home how dangerous this poo poo is to the average sheeple. 2 more weeks. Hopper fucked around with this message at 08:01 on Apr 15, 2020 |
# ? Apr 15, 2020 06:08 |
|
Hopper posted:This is Europe, not the US, when has anybody ever been sued and properly punished for something that wasn't a criminal activity? *laughs in unconstitutional Hartz IV sanctions*
|
# ? Apr 15, 2020 07:55 |
|
Did anybody get punished?
|
# ? Apr 15, 2020 08:02 |
|
The Hartz IV Empfänger.
|
# ? Apr 15, 2020 08:35 |
|
Smirr posted:I had a quiet, internal meltdown at the FT graphs a couple of days ago, so I guess I'll turn that into a very calm external post: [...] For a meltdown, that was pretty civilised. But why do you suspect the graphs to be intentionally misleading?
|
# ? Apr 15, 2020 09:13 |
|
Randler posted:The Hartz IV Empfänger. And this justice was Done. - SPD Sprichwort
|
# ? Apr 15, 2020 09:18 |
|
Hope you're safe. Randler posted:I adjusted the FT curve for a more realistic assessment of the situation. It is a joy and beauty! While i honestly appreciate y'all's musing about this, now that i've had sleep, and agree with your concerns in re the average population of non-number people. I'd like to point out that the readership there is the financial times, an audience that can be expected to read the descriptions of the graphs, so they're kinda optimizing for that. For me the features of the graphs are: - shows whether a country is actively worsening (up), at least not immediately worsening (flat), keeping somewhat of a grip (down) - provides a goal (0) - dampens outliers (7-days rolling average) - shows a readably behavior curve for countries of all sizes (log) That said, i would be curious what a version with normalized values instead of log looks like.
|
# ? Apr 15, 2020 09:27 |
|
Paternoster elevator would have been a lot more distanced and only a little slower for everyone, maybe they'll have a resurgence.
My Lovely Horse fucked around with this message at 09:37 on Apr 15, 2020 |
# ? Apr 15, 2020 09:35 |
3 supermarkets around the corner from our apartment and not a single one of them has toilet paper.
|
|
# ? Apr 15, 2020 09:59 |
|
DTurtle posted:Oh right. Cool, thanks.
|
# ? Apr 15, 2020 10:10 |
|
Einbauschrank posted:For a meltdown, that was pretty civilised. But why do you suspect the graphs to be intentionally misleading? Because I suspect that the percentage of death-worshipping "open it all up already" shouters will be pretty high among FT readers when comparing newspapers. And these graphs have the unfortunate property that the curves will flatten sooner than they would on a 1) log scale plot of total cases, 2) linear scale plot of new cases. Like, they will, for any given curve. Mithaldu posted:While i honestly appreciate y'all's musing about this, now that i've had sleep, and agree with your concerns in re the average population of non-number people. I'd like to point out that the readership there is the financial times, an audience that can be expected to read the descriptions of the graphs, so they're kinda optimizing for that. I guess my fundamental problem is that most people probably heard of "flatten the curve" in the context of a plot with a line that goes from 0 to x exponentially, then down to 0 exponentially, overlaid with a flat line indicating hospital capacity. That is either a plot of 1) total number of active cases, cumulative, (perhaps just those requiring hospitalization) and total hospital capacity, or 2) number of new cases per day (not cumulative) and some sort of measure of per-day hospital capacity (which has to take into account the absolute capacity, but also turnover by release or deaths, which is determined by the disease, previous surges, etc.). I just realized that my issue is actually mostly independent of the log scale. The FT graphs are type 2, where it would be quite tricky to determine where the (almost certainly not flat) line for hospital capacity is to go (and thus which curve counts as flattened just enough). On the other hand, all those example graphs for the concept of flattening the curve are all type 1, I'm pretty sure, or at least the straight line on it explicitly has the simple interpretation of "this is how many beds there are, total", which requires the number of cases to be the number of active cases. I suspect that there are quite a few people who see a peaking curve on the FT graphs and think "well, as long as this maximum was below the flat line of number of beds, we're through the worst now". But non-cumulative new cases per day require a complicated concept of hospital capacity, so we just don't know that. Basically, right now I would prefer any plot of covid cases as a curve (in the sense of 0 -> x -> 0) to allow for the straightforward capacity interpretation, since that's the most important thing the lockdown was justified with. The log scale is just a secondary thing that makes linear growth look like no growth at a glance. Yes, carefully reading the graph will reveal that misinterpretation, but it will make things look better at first sight.
|
# ? Apr 15, 2020 10:29 |
|
Hopper posted:This is Europe, not the US, when has anybody ever been sued and properly punished for something that wasn't a criminal activity? Man versucht es zumindest: https://www.dw.com/de/corona-versuchte-vertuschung-in-tirol/a-53100314?maca=de-VEU_Focus-Artikel_DE_Europa-29099-xml-media und kriminelle Aktivität könnte auch kommen (aber da ÖVP in Tirol herrscht und ÖVP in Wien herrscht, wird da nix rauskommen). https://www.faz.net/aktuell/politik/ausland/tirol-will-aufklaerung-in-corona-krise-der-filz-von-ischgl-16694930.html
|
# ? Apr 15, 2020 10:49 |
|
Auch ein sehr interessanter Artikel über Hendrik Streek, ein Virologe, der anscheinend über gefärbte Aussagen und Studien etwas Politik machen will oder zumindest den "End Shutdown"-Leuten zuarbeiten will: https://www.riffreporter.de/corona-virus/corona-streeck-heinsberg-pandemie-exit-laschet/
|
# ? Apr 15, 2020 10:54 |
|
Mithaldu posted:Hope you're safe. Yeah thanks, should be. I literally only have to walk 250 meters to the supermarket, if that. And this whenever I deem it too busy I just turn around and come bakc at some other time. Today was the odd one out with that lady squeezing past, usually there is enough room around me when I wear the mask. I do it because a) people keep more distance and b) I won't know immediately once I get the roni, so I don't want to put others at risk. And I've traveled to Thailand, Taiwan and Japan, that taught me that masks in public are nothing special, so I just go along with it. I just wish we had some of their stylish reusable ones that make you look like ninjas...
|
# ? Apr 15, 2020 10:59 |
|
I still have only one pair from my Arbeitgeber with sick flower/batic print that make me look like an extremely alternative ninja. Also, I'm watching the briefing of the "Corona Kabinett", and they're being extrem durch die Blume, but it sounds like it's pretty safe to assume that they won't be recommending lifting of the Maßnahmen, or even considering lifting the Maßnahmen before early May. https://www.youtube.com/watch?v=EZpHFe1QJIY @TAG DER ENTSCHEIDUNG /edit: Misspoke there, since "the Maßnahmen" are largely Sache der Länder, so emphasis on "recommending" /edit: Grenzkontrollen and Grenz-bezogene Einschränkungen are extended until May 4th (be with you, etc.) Duzzy Funlop fucked around with this message at 11:36 on Apr 15, 2020 |
# ? Apr 15, 2020 11:27 |
|
This got posted by an verbeamtetet Lehrerin that I know : https://editionf.com/die-lockerungen-der-corona-massnahmen-duerfen-nicht-auf-dem-ruecken-von-eltern-ausgetragen-werden/ lol that Milchmädchenrechnung quote:Der Tag hat auch für Eltern in Corona-Zeiten nur 24 Stunden. Good thing it wasn't me who got her preggers lol
|
# ? Apr 15, 2020 12:31 |
|
You don't win the prestigious Cleanest House in Germany award without putting in some work.
|
# ? Apr 15, 2020 12:40 |
|
ephex posted:This got posted by an verbeamtetet Lehrerin that I know : Force all parents into kurzarbeit. I mean it makes sense, if they don't need to pay for Kinderbetreuung anymore, they have the gap in money to spare. cant cook creole bream fucked around with this message at 12:45 on Apr 15, 2020 |
# ? Apr 15, 2020 12:41 |
|
ephex posted:This got posted by an verbeamtetet Lehrerin that I know : maybe stop showering for 45 minutes per parent/day, jfc
|
# ? Apr 15, 2020 12:43 |
|
ephex posted:This got posted by an verbeamtetet Lehrerin that I know : Hope she isn't a math teacher
|
# ? Apr 15, 2020 12:45 |
|
GABA ghoul posted:Hope she isn't a math teacher lmao of course she is
|
# ? Apr 15, 2020 12:46 |
|
ephex posted:This got posted by an verbeamtetet Lehrerin that I know : It completely rules how many people pretty openly hate their children and don't actually want to be around them for any substantial time, and their takeaway is not "wow, the people that I foist my brats on are braver than the troops and criminally underpaid", it's "I should get their money for the time being", lol. Sure, there are all sorts of interesting economic ramifications here too, but I'll stick with "eat poo poo, childhavers"
|
# ? Apr 15, 2020 12:51 |
|
Smirr posted:It completely rules how many people pretty openly hate their children and don't actually want to be around them for any substantial time, and their takeaway is not "wow, the people that I foist my brats on are braver than the troops and criminally underpaid", it's "I should get their money for the time being", lol. Sure, there are all sorts of interesting economic ramifications here too, but I'll stick with "eat poo poo, childhavers" I love this post
|
# ? Apr 15, 2020 12:53 |
|
Also, at the risk of sounding dismissive of the plight of parents, 14 uninterupted hours of watching a child each day seems excessive. Do these kids only get 2 waking hours without their parents continously staring at them? Surely, during that time there would be a moment to take out your phone and check your corona news.
|
# ? Apr 15, 2020 12:55 |
|
You know, i absolutely did not make the connection between "flatten the (currently sick patients) curve", a (new cases per day) curve in a financial mag that is going down, and the possible misunderstandings that lead to financial people yelling "reopening is ok". Good point.
Neat, except for the "i've traveled to asia" bit your situation is almost exactly mine (tho i have that by proxy via the hong kong circuit). You could get black cloth masks from amazon and doodle no them. Or check etsy, lotsa people are selling colorful reusables for 10 a piece. ephex posted:This got posted by an verbeamtetet Lehrerin that I know : quote:In der Grundschule sind es 2,8 Millionen Schüler*innen in Klassengrößen von rund 20 Kindern. I thinkt he word "rund" is doing a lot of work there.
|
# ? Apr 15, 2020 13:24 |
|
Umm, does anyone else have issues with this fit (and the corresponding result that with >50% working from home jobs people will turn into zombies). Cause it seems really really bad, a couple of outliers on the left really screw with the LS fit. ] Ika fucked around with this message at 13:28 on Apr 15, 2020 |
# ? Apr 15, 2020 13:26 |
|
quote:Zudem waren die meisten Familien vorher auch schon sehr effizient organisiert.
|
# ? Apr 15, 2020 13:26 |
|
38%-46% is a very small range and the whole thing smells of noise to me, but i'm just a programmer and i think Smirr might have strong opinions.
|
# ? Apr 15, 2020 13:28 |
|
|
# ? Jun 8, 2024 07:41 |
|
ephex posted:This got posted by an verbeamtetet Lehrerin that I know : You don't have to read past the 16 hours sleep. No parent gets 8 hours of sleep unless their kids are teenagers or older and then the rest is bullshist because teenagers don't need 14hrs of attention, they'll just watch twitch or whatever. But yeah 1.5h Körperpflege every day day? Wtf?
|
# ? Apr 15, 2020 13:29 |