|
The new Global Cooling thread.
Server: Shiva
Game: FFXI
Posts: 20130
By Shiva.Nikolce 2014-06-24 13:33:36
Anyone who has handled raw data
Should wash their hands thoroughly and also wash any and all of the surfaces that may have come into contact with the raw data to prevent cross contamination.
Cerberus.Pleebo
Server: Cerberus
Game: FFXI
Posts: 9720
By Cerberus.Pleebo 2014-06-24 13:59:17
None.
So, tell me, why were there additional "scrubbing" of data done to the raw data after the obvious errors were taken out?
Why were there so many errors to begin with?
One would expect to see that the unadjusted and adjusted data to be roughly the same, but that is not the case with this. What additional "errors" were found that is not reported, and are we sure that the "errors" found and removed were actual "errors" listed on Tables 1 and 2? No, specifically. Tell me specifically what you find faulty about their data processing. Pick any of their bias corrections and explain. Bias correction being necessary for a very large, very heterogeneous data set isn't even up for debate. Even Saevel's *** WUWT link acknowledges as much.
Altima couldn't seem to find the words so you'll have to take his place as someone who obviously knows what they're talking about when it comes to this analysis. So please, fill me in.
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:01:28
None.
So, tell me, why were there additional "scrubbing" of data done to the raw data after the obvious errors were taken out?
Why were there so many errors to begin with?
One would expect to see that the unadjusted and adjusted data to be roughly the same, but that is not the case with this. What additional "errors" were found that is not reported, and are we sure that the "errors" found and removed were actual "errors" listed on Tables 1 and 2? No, specifically. Tell me specifically what you find faulty about their data processing. Pick any of their bias corrections and explain. Bias correction being necessary for a very large, very heterogeneous data set isn't even up for debate. Even Saevel's *** WUWT link acknowledges as much.
Altima couldn't seem to find the words so you'll have to take his place as someone who obviously knows what they're talking about when it comes to this analysis. So please, fill me in. Thank you for acknowledging that I understand the data and methods better than you.
I would love to explain, but I'm afraid that it might be a little too much for you, and I don't want to waste either mine or your time.
Cerberus.Pleebo
Server: Cerberus
Game: FFXI
Posts: 9720
By Cerberus.Pleebo 2014-06-24 14:04:24
HAHAHAHHAHAHAHAHAHAHAHHAHAHHAHAHHAHAHA
Excellent cop out, my good man. You got a very audible guffaw out of me.
Now if we could only get Bill Herp The Science Derp to stop making threads about things he doesn't understand we'd be making progress.
[+]
Bahamut.Milamber
Server: Bahamut
Game: FFXI
Posts: 3691
By Bahamut.Milamber 2014-06-24 14:06:00
None.
So, tell me, why were there additional "scrubbing" of data done to the raw data after the obvious errors were taken out?
Why were there so many errors to begin with?
One would expect to see that the unadjusted and adjusted data to be roughly the same, but that is not the case with this. What additional "errors" were found that is not reported, and are we sure that the "errors" found and removed were actual "errors" listed on Tables 1 and 2? No, specifically. Tell me specifically what you find faulty about their data processing. Pick any of their bias corrections and explain. Bias correction being necessary for a very large, very heterogeneous data set isn't even up for debate. Even Saevel's *** WUWT link acknowledges as much.
Altima couldn't seem to find the words so you'll have to take his place as someone who obviously knows what they're talking about when it comes to this analysis. So please, fill me in. Thank you for acknowledging that I understand the data and methods better than you.
I would love to explain, but I'm afraid that it might be a little too much for you, and I don't want to waste either mine or your time. That isn't anywhere in the realm of what he said. Unless you are purposefully misunderstanding in order to flirt, then sail on sailor!
[+]
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:06:55
HAHAHAHHAHAHAHAHAHAHAHHAHAHHAHAHHAHAHA
Excellent cop out, my good man. You got a very audible guffaw out of me.
There is not enough time in the whole timespan to educate you in methods and meanings of analytic reporting.
Especially in regards to this issue.
I mean, it would take you years to understand why too much data scrubbing is not good.
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:08:05
Because it isn't a perfect world, and there are quite a number of non-obvious ways for sensors to either fail or report non-true data, and no-one so far has spent the money to be properly redundant at any given site (by general appearances)?
I wouldn't necessarily expect raw/corrected data to appear the same; it really, really depends on potential error sources. Real data is not necessarily easy or cheap to acquire in large, distributed systems. The main issue is, if the data itself is so bad, what point would be to use it if it requires so much alterations to the data in order for it to be "usable"?
Cerberus.Pleebo
Server: Cerberus
Game: FFXI
Posts: 9720
By Cerberus.Pleebo 2014-06-24 14:08:14
I said pick one. Any one of them. If you want to critique methodology, then do it.
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:11:28
I said pick one. Any one of them. If you want to critique methodology, then do it. I don't have an issue in the methodology.
I have an issue with the overall product.
The whole purpose of adjusting the data that takes out errors is to give better, usable data.
But if the data was so horrible that, after the alterations, it produces an entire different set of data, then what is the point of the original data in the first place? That signifies that the method of obtaining the data is flawed and the data itself should never be used, altered or raw, because it is inherently incorrect.
VIP
Server: Odin
Game: FFXI
Posts: 9534
By Odin.Jassik 2014-06-24 14:14:02
"I can't find any reason to disagree besides not liking the narrative"
[+]
Bahamut.Milamber
Server: Bahamut
Game: FFXI
Posts: 3691
By Bahamut.Milamber 2014-06-24 14:15:21
Because it isn't a perfect world, and there are quite a number of non-obvious ways for sensors to either fail or report non-true data, and no-one so far has spent the money to be properly redundant at any given site (by general appearances)?
I wouldn't necessarily expect raw/corrected data to appear the same; it really, really depends on potential error sources. Real data is not necessarily easy or cheap to acquire in large, distributed systems. The main issue is, if the data itself is so bad, what point would be to use it if it requires so much alterations to the data in order for it to be "usable"?
Physical systems can have constant, semi-constant, or otherwise compensateable errors. What can be bad on its own can be corrected for based on the type of error (if known), however some cannort be recovered or compensated for; these samples cannot be used.
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:15:39
"I can't find any reason to disagree besides not liking the narrative" Wait, you are saying that it is ok to make ***up and run with it?
VIP
Server: Odin
Game: FFXI
Posts: 9534
By Odin.Jassik 2014-06-24 14:19:10
"I can't find any reason to disagree besides not liking the narrative" Wait, you are saying that it is ok to make ***up and run with it?
I think it's not ok to claim credulity in matters far outside your capabilities. It's not honest to expect the world to slow down for you because you don't like reality.
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:19:13
Because it isn't a perfect world, and there are quite a number of non-obvious ways for sensors to either fail or report non-true data, and no-one so far has spent the money to be properly redundant at any given site (by general appearances)?
I wouldn't necessarily expect raw/corrected data to appear the same; it really, really depends on potential error sources. Real data is not necessarily easy or cheap to acquire in large, distributed systems. The main issue is, if the data itself is so bad, what point would be to use it if it requires so much alterations to the data in order for it to be "usable"?
Physical systems can have constant, semi-constant, or otherwise compensateable errors. What can be bad on its own can be corrected for based on the type of error (if known), however some cannort be recovered or compensated for; these samples cannot be used. There are limits of compensation though, before the whole pool is tainted.
If it was decades before the errors in the record-keeping was discovered, then the real issue is the date prior to the correction of the methods of record-keeping. To use those errors knowingly not only taints the actual study being performed, but also commits "intellectual dishonesty" or "intellectual fraud," which in this case, seems to be committed by the recording/reporting agencies (not the scientists themselves, unless they knew that the data is incorrect and still used them anyway).
Ever heard of the concept: "garbage in, garbage out"?
Cerberus.Pleebo
Server: Cerberus
Game: FFXI
Posts: 9720
By Cerberus.Pleebo 2014-06-24 14:19:29
I don't have an issue in the methodology.
I have an issue with the overall product.
The whole purpose of adjusting the data that takes out errors is to give better, usable data.
But if the data was so horrible that, after the alterations, it produces an entire different set of data, then what is the point of the original data in the first place? That signifies that the method of obtaining the data is flawed and the data itself should never be used, altered or raw, because it is inherently incorrect. Quality assurance is part of it. Stop stalling and present something with substance. What are your specific objections with the bias correction and, since you brought it up, which systematic errors do you consider most fatal to the further use of the data set?
[+]
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:20:20
"I can't find any reason to disagree besides not liking the narrative" Wait, you are saying that it is ok to make ***up and run with it?
I think it's not ok to claim credulity in matters far outside your capabilities. It's not honest to expect the world to slow down for you because you don't like reality. I think it is not ok to claim credibility in matter where a so-called expert in the field knowingly reports errors.
VIP
Server: Odin
Game: FFXI
Posts: 9534
By Odin.Jassik 2014-06-24 14:20:23
Ever heard of the concept: "garbage in, garbage out"?
Ever hear of the concept: "talking out your ***?"
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:21:55
I don't have an issue in the methodology.
I have an issue with the overall product.
The whole purpose of adjusting the data that takes out errors is to give better, usable data.
But if the data was so horrible that, after the alterations, it produces an entire different set of data, then what is the point of the original data in the first place? That signifies that the method of obtaining the data is flawed and the data itself should never be used, altered or raw, because it is inherently incorrect. Quality assurance is part of it. Stop stalling and present something with substance. What are your specific objections with the bias correction and, since you brought it up, which systematic errors do you consider most fatal to the further use of the data set? Lets see, I answered your same question twice, with the same answer, and you still asked the same question again.
I think that qualifies as lack of reading comprehension. Which is actually common with you.
So, thank you for proving once again your idiocy.
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:22:38
Ever heard of the concept: "garbage in, garbage out"?
Ever hear of the concept: "talking out your ***?" No, but thank you for the examples you provided in your posts.
VIP
Server: Odin
Game: FFXI
Posts: 9534
By Odin.Jassik 2014-06-24 14:25:56
Ever heard of the concept: "garbage in, garbage out"?
Ever hear of the concept: "talking out your ***?" No, but thank you for the examples you provided in your posts.
Your posts are the internet forum equivalent of "I know you are but what am I?" Learn some humility, you aren't in any way qualified to question the findings of experts and even less qualified to participate in a scholarly debate.
Cerberus.Pleebo
Server: Cerberus
Game: FFXI
Posts: 9720
By Cerberus.Pleebo 2014-06-24 14:26:52
No, you gave me the Sarah Palin answer of "Uh... all of them". I want a specific answer. This is how scientists do it. We're all waiting.
Bahamut.Milamber
Server: Bahamut
Game: FFXI
Posts: 3691
By Bahamut.Milamber 2014-06-24 14:28:06
Because it isn't a perfect world, and there are quite a number of non-obvious ways for sensors to either fail or report non-true data, and no-one so far has spent the money to be properly redundant at any given site (by general appearances)?
I wouldn't necessarily expect raw/corrected data to appear the same; it really, really depends on potential error sources. Real data is not necessarily easy or cheap to acquire in large, distributed systems. The main issue is, if the data itself is so bad, what point would be to use it if it requires so much alterations to the data in order for it to be "usable"?
Physical systems can have constant, semi-constant, or otherwise compensateable errors. What can be bad on its own can be corrected for based on the type of error (if known), however some cannort be recovered or compensated for; these samples cannot be used. There are limits of compensation though, before the whole pool is tainted.
If it was decades before the errors in the record-keeping was discovered, then the real issue is the date prior to the correction of the methods of record-keeping. To use those errors knowingly not only taints the actual study being performed, but also commits "intellectual dishonesty" or "intellectual fraud," which in this case, seems to be committed by the recording/reporting agencies (not the scientists themselves, unless they knew that the data is incorrect and still used them anyway).
Ever heard of the concept: "garbage in, garbage out"? You can correct different sections of data different ways, for different reasons; that is essential to do when you have different sources of different types. You generally are required to take the worst error margin of sampling, but that can be different based on how much the error for that dataset contributes to the overall sample.
Yes, I am extremely familiar with GIGO. That's generally why you even perform this in the first place, to even determine if you can draw any conclusions from the data, or to what extent.
[+]
Server: Shiva
Game: FFXI
Posts: 20130
By Shiva.Nikolce 2014-06-24 14:32:32
I'm not. I lost interest a long time ago.
But we can scrub the data so it still looks like everyone is interested!
[+]
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:38:46
You can correct different sections of data different ways, for different reasons; that is essential to do when you have different sources of different types. You generally are required to take the worst error margin of sampling, but that can be different based on how much the error for that dataset contributes to the overall sample.
Yes, I am extremely familiar with GIGO. That's generally why you even perform this in the first place, to even determine if you can draw any conclusions from the data, or to what extent. So let me ask you this, since it seems like we have come to an impasse.
How much data would have to be altered before the data set to be considered useless?
I think that a 5% error rate would signify issues in regards to the record keeping of the data, and anything above that would throw the entire set into question of feasibility. 6% error rate should automatically disqualify the data from usability.
Of course, the data we are talking about in question is greater than 6%.
What are your thoughts in that?
VIP
Server: Odin
Game: FFXI
Posts: 9534
By Odin.Jassik 2014-06-24 14:44:09
You can correct different sections of data different ways, for different reasons; that is essential to do when you have different sources of different types. You generally are required to take the worst error margin of sampling, but that can be different based on how much the error for that dataset contributes to the overall sample.
Yes, I am extremely familiar with GIGO. That's generally why you even perform this in the first place, to even determine if you can draw any conclusions from the data, or to what extent. So let me ask you this, since it seems like we have come to an impasse.
How much data would have to be altered before the data set to be considered useless?
I think that a 5% error rate would signify issues in regards to the record keeping of the data, and anything above that would throw the entire set into question of feasibility. 6% error rate should automatically disqualify the data from usability.
Of course, the data we are talking about in question is greater than 6%.
What are your thoughts in that?
You seem to have unrealistic expectations of data accuracy spanning the types and timeframe that climate data does. Also very little understanding of how data is analyzed.
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:50:29
You seem to have unrealistic expectations of data accuracy spanning the types and timeframe that climate data does. Wait, it's more advanced than "The temperature at XXX was XX degrees as the high, recorded at X:XX, XX degrees as the low, recorded at X:XX, the barometric pressure was XXX, humidity was at XXX, it was <insert weather> for XX hours, <insert weather> for XX hours, (so on), and XX amount of rain was collected"?
You know, something that is recorded at every airport in the country on a daily basis (at the very least).
Oh, I forgot, there's "climate speech" included in that data that only "climate scientists" would know about.
Also very little understanding of how data is analyzed. No, that's you bub.
Cerberus.Pleebo
Server: Cerberus
Game: FFXI
Posts: 9720
By Cerberus.Pleebo 2014-06-24 14:52:09
The simple answer to that question is yes. Yes it is more complex than a weather station at the *** airport. Now stop trying to steer the conversation away from my questions.
Ragnarok.Yatenkou
Server: Ragnarok
Game: FFXI
Posts: 319
By Ragnarok.Yatenkou 2014-06-24 14:54:46
Idc what it's called, IM TIRED OF THESE STORMS!!!!
[+]
Server: Asura
Game: FFXI
Posts: 34187
By Asura.Kingnobody 2014-06-24 14:57:31
Ragnarok.Yatenkou said: »Idc what it's called, IM TIRED OF THESE STORMS!!!! You forgot to blame Bush for the storms.
Shame on you!
Server: Fenrir
Game: FFXI
Posts: 127
By Fenrir.Squintik 2014-06-24 15:01:18
Ragnarok.Yatenkou said: »Idc what it's called, IM TIRED OF THESE STORMS!!!!
These storms are at least cooling things down a bit, so I don't mind them so much (except when they kill my internet connection). I'm just tired of this insane heat.
Turns out the US has been cooling sense the 1930's when you look at charts/graphs that have not been "altered". Looks like I'm not the only one that has a problem with fudged data.
http://stevengoddard.wordpress.com/data-tampering-at-ushcngiss/
|
|