|
morallyobjected posted:OTOH maybe we should be glad some of the fields aren't trying to interact with each other
|
#
?
Sep 11, 2020 08:37
|
|
|
|
| # ? May 27, 2024 14:05 |
|
|
All good science starts with a basic question: What do I believe and how do I create a dataset to prove it?
|
|
#
?
Sep 11, 2020 14:37
|
|
|
Outrail posted:All good science starts with a basic question: What do I believe and how do I create a dataset to prove it? This same person later tries to argue "If we don't publish the numbers about how bad things are, things won't be as bad!"
|
|
#
?
Sep 11, 2020 17:51
|
|
|
Teriyaki Hairpiece posted:This same person later tries to argue "If we don't publish the numbers about how bad things are, things won't be as bad!" Actually, if we remove the outliers outside our hypothetical norm the dataset looks pretty good
|
|
#
?
Sep 11, 2020 18:00
|
|
|
Just sort your dependent and independent values separately you cowards. That way you always get good fits.
|
|
#
?
Sep 11, 2020 18:12
|
|
|
Outrail posted:Actually, if we remove the outliers outside our hypothetical norm the dataset looks pretty good Winsorize your data to look normal
|
|
#
?
Sep 11, 2020 18:53
|
|
|
BonHair posted:Just sort your dependent and independent values separately you cowards. That way you always get good fits. All published graphs converge to x=y
|
|
#
?
Sep 11, 2020 22:55
|
|
|
 Very helpful, thank you.
|
|
#
?
Sep 12, 2020 17:08
|
|
|
FreeKillB posted:
voting participation is the new dick size. "Yeah my participation is twice the size of an average one"
|
|
#
?
Sep 13, 2020 01:18
|
|
|
Outrail posted:Actually, if we remove the outliers outside our hypothetical norm the dataset looks pretty good Sometimes when you massage the numbers you get a happy ending
|
|
#
?
Sep 13, 2020 04:29
|
|
|
|
|
#
?
Sep 18, 2020 23:29
|
|
|
Do the remaining 1% think we're multi-track drifting?
|
|
#
?
Sep 19, 2020 10:10
|
|
|
A lesson in being clear what data you're leaving out, here: https://twitter.com/simongerman600/status/1307983243719913473
|
|
#
?
Sep 23, 2020 12:37
|
|
|
Angepain posted:A lesson in being clear what data you're leaving out, here: That is interesting to see the spread westward and the gold rush in California. Also, as is commonly known, no humans lived in the Americas prior to the Europeans discovering it.
|
|
#
?
Sep 23, 2020 13:09
|
|
|
|
|
#
?
Sep 23, 2020 13:11
|
|
|
What's the chance it doesn't count black people up until like the 1960s?
|
|
#
?
Sep 23, 2020 13:34
|
|
|
It counts three‐fifths of them.
|
|
#
?
Sep 23, 2020 13:41
|
|
|
Platystemon posted:It counts three‐fifths of them. Ni....not nice.
|
|
#
?
Sep 23, 2020 14:15
|
|
|
jjack229 posted:Also, as is commonly known, no humans lived in the Americas prior to the Europeans discovering it. It's equally commonly known that the census data it cites made sure that they measured the population of tribes they didn't even know existed yet. This map does a whole bunch of very loaded things (and the census office didn't do density maps until 1860) but that's not one of them.
|
|
#
?
Sep 23, 2020 14:54
|
|
|
Is that squares per mile? Squids per mile? [Pop. Density]^2 per mile?
|
|
#
?
Sep 23, 2020 15:50
|
|
|
Sighence posted:It's equally commonly known that the census data it cites made sure that they measured the population of tribes they didn't even know existed yet. It kind of is one of them, in that while you could technically conclude that from seeing the small text saying it's census data and considering the likely population of the census, the way the data is presented doesn't make clear that there are in fact people there. The map isn't technically lying on that front, but it is misleading in a way that reinforces some forms of ignorance. Also worth nothing that this map isn't complete or precise either, from the comments in that tweet the guy who made it acknowledges a work in progress and it's only part of the data. https://twitter.com/RWArchaeology/status/1308447051752103938
|
|
#
?
Sep 23, 2020 15:58
|
|
|
Angepain posted:Also worth nothing that this map isn't complete or precise either, from the comments in that tweet the guy who made it acknowledges a work in progress and it's only part of the data.
|
|
#
?
Sep 23, 2020 16:44
|
|
|
BonHair posted:What's the chance it doesn't count black people up until like the 1960s? In about the 1880s and 1890s, you can see population density start to rise in That Famous Cretaceous Coastline, so the odds are that the census probably didn't count black people properly until after the civil war..
|
|
#
?
Sep 24, 2020 05:46
|
|
|
FreeKillB posted:
This one pisses me off because apparently my voting record is "average", even though I've voted in every single election (primaries, midterms, random school referendums, etc) since I turned 18.
|
|
#
?
Sep 24, 2020 06:48
|
|
|
Raldikuk posted:This one pisses me off because apparently my voting record is "average", even though I've voted in every single election (primaries, midterms, random school referendums, etc) since I turned 18. that's average for you tho
|
|
#
?
Sep 24, 2020 12:43
|
|
|
Raldikuk posted:This one pisses me off because apparently my voting record is "average", even though I've voted in every single election (primaries, midterms, random school referendums, etc) since I turned 18. I want to guess that it's based on a metric that's scored in the Democratic voter database, and I don't think anyone under the age of 70 gets a good score, no matter how regularly they vote.
|
|
#
?
Sep 24, 2020 17:45
|
|
|
Raldikuk posted:This one pisses me off because apparently my voting record is "average", even though I've voted in every single election (primaries, midterms, random school referendums, etc) since I turned 18.
|
|
#
?
Sep 25, 2020 07:09
|
|
|
I mean it would have to be, since some people don't vote, and it's hard to overvote.
|
|
#
?
Sep 25, 2020 08:18
|
|
|
Golbez posted:I mean it would have to be, since some people don't vote, and it's hard to overvote. There is some disagreement on this point...
|
|
#
?
Sep 25, 2020 08:31
|
|
|
https://twitter.com/baumard_nicolas/status/1308715609795039232
|
|
#
?
Sep 25, 2020 08:58
|
|
|
quote:automatically generate trustworthiness evaluations for the facial action units of Oblivion characters
|
|
#
?
Sep 25, 2020 09:06
|
|
|
Dr. Arbitrary posted:I want to guess that it's based on a metric that's scored in the Democratic voter database, and I don't think anyone under the age of 70 gets a good score, no matter how regularly they vote. lol i bet their "average" voter has voted in at least 10 presidential elections
|
|
#
?
Sep 25, 2020 09:57
|
|
|
|
|
#
?
Sep 25, 2020 11:31
|
|
|
Did they publish it in quilette
|
|
#
?
Sep 25, 2020 12:16
|
|
|
We showed a computer smiling vs frowny faces and told it happy faces were better.
|
|
#
?
Sep 25, 2020 12:22
|
|
|
90s Cringe Rock posted:Did they publish it in quilette doubt it, both faces are white
|
|
#
?
Sep 25, 2020 13:48
|
|
|
But you guys they compared the algorithm to results obtained from human surveys and it matched! Which they apparently dont' realize is just another way of saying they've automated the obvious euro-centric bias of their focus group, by training an AI on historical euro-centric preferences. 
|
|
#
?
Sep 25, 2020 14:31
|
|
|
Also, they're using paintings as data. I'm not an art historian but I don't think it's true that all paintings are photo-realistic and there have been no variations in style and technique in European art over 500 years, which is what they seem to be assuming.
|
|
#
?
Sep 25, 2020 14:42
|
|
|
Well supercomputers didn't solve all our problems, but at least now we can prove phrenology works.
|
|
#
?
Sep 25, 2020 14:52
|
|
|
|
| # ? May 27, 2024 14:05 |
|
|
steinrokkan posted:Well supercomputers didn't solve all our problems, but at least now we can prove phrenology works. Super computers + Phrenology = Super Phrenology! I call this Super Phrenology. And THIS is Super Phrenology II!
|
|
#
?
Sep 25, 2020 15:08
|
|