The Glade 4.0

It's funny, the biggest problem I have with the map is something no one else seems to have noticed. And now that I've thought about it a bit more, I think it explains the east/west divide perfectly.

All their data is aggregated to the county level. When you zoom in on the map, you'll see dots on the map that appear to be little pockets of racism. As you zoom in more, these pockets get smaller. They're really just the geographic center of a county. The size of the blue/red blob has nothing to do with the size of the county. When you start to zoom out, you see a bunch of little blue blobs in the eastern US, all of similar size. Counties in the east are smaller and more densely packed together than in the west, so you see more of these blobs there and it makes the east look really bad. It's really an issue with the way they are displaying their map on Google Maps, not the data itself.

If you look at this map of the counties in the US, you'll see it lines up almost perfectly with this east/west divide.



They really need to have a map of all the counties like the one above, with each county shaded according to their data. Then it would give a much clearer picture of the data and I suspect this east/west divide would vanish.

That's truer than you think ...

_________________

What, did that come off as sarcastic?

It came off quite sarcastic, actually. And I'm not belittling the notion that the US is full of small-minded people with irrational fears and social flaws, in the slightest. The US is a lot more homogeneous than people think.

_________________

Fair enough. I'll remove my response.

_________________
“The duty of a patriot is to protect his country from its government.” - Thomas Paine

Interesting.

Lenas:

If you're curious, it comes across as sarcastic because of the diction and idiomatic choices in your statement. As for the substantives, I'd point out that ethnic qualifiers are by-and-large irrelevant in discussions of diversity, ability, and various behavior-set analysis that our media and politicians like to discuss. African Americans, as a racial group for instance, may not suffer from institutional racism in the way Monte used to argue, for instance. The situation might be far less transparent and tangible. It could be, for instance, that income and generational spending habits have more to do with the situation than race. It is true, however, that people tend to group far more rigidly along income lines than they do along ethnic or gender or sexuality lines. In fact, that's largely why Boulder County, Colorado is the poster child for both sides of the diversity debate.

_________________

Just goes to show, that the further East you go, the more uptight and assholish the residents get. Go far enough West, and people tend to get more laid back and relaxed. Just look at the differences between Hawaii and New York City.

_________________

Holy shitsnacks!

On this seaboard, we like to point out that the first step (of 12) in getting out of California lies with remembering you don't have to keep up with the Jones.

_________________

While I was marginally aware of this phenomena, It didn't occur to me that it might apply.

_________________
I prefer to think of them as "Fighting evil in another dimension"

So here's the key for me, basically (ordered sequentially, not by order importance):

1) There are a TON of biases and confounding variables in this map.
2) According to the authors, (some of) these are known biases
3) They chose to publish the findings anyway, despite the known biases

Therefore:

4) The findings are not valid due to intentional disregard for known biases.



Now, if you'd like me to go into greater depth as to why this is (and rightly should be) the case, I can.

_________________
“The duty of a patriot is to protect his country from its government.” - Thomas Paine

This isn't even data. It's a mountain of useless information.

_________________
Buckle your pants or they might fall down.

All collected social information is biased in one way or another.

That doesn't make it entirely useless. It just increases the possible causes for the data to show what it shows. Rather than allow a conclusion, it simply shows interesting possibilities.

_________________
Well Ali Baba had them forty thieves, Scheherezade had a thousand tales
But master you in luck 'cause up your sleeves you got a brand of magic never fails...
...Mister Aladdin, sir, What will your pleasure be?
Let me take your order, Jot it down -You ain't never had a friend like me
█ ♣ █

In this case, it is entirely useless. The rest of your statement is entirely correct.

_________________
“The duty of a patriot is to protect his country from its government.” - Thomas Paine

And yet, biased social information polls are able to predict the results of an election with a reasonable level of accuracy. Biased social information is used to help companies choose product lines, predict customer volumes (and staff/stock accordingly), etc.

The information is useful. it just cannot, on its own, show causal relationships, just corellations.

_________________
Well Ali Baba had them forty thieves, Scheherezade had a thousand tales
But master you in luck 'cause up your sleeves you got a brand of magic never fails...
...Mister Aladdin, sir, What will your pleasure be?
Let me take your order, Jot it down -You ain't never had a friend like me
█ ♣ █

Again, like with Amanar, if that's what you want to believe, fine. But you're wrong. This data, and this map, show you effectively nothing.


The other items you mention are false analogies. Essentially, what your core point is correct: not all variables can be controlled. This does not mean that all information is valid, or even useful. Some data has so many variables, and so many biases, as to be wholly without worth.

This is (essentially) the case here.

_________________
“The duty of a patriot is to protect his country from its government.” - Thomas Paine

To give those of you who think this has valuable information an idea of how bad the results of the study are ...

Apparently the University of Georgia is so racist that it turns two states (Tennessee and Georgia) red. The most I look at this, the most I see a gross expression of some researcher's confirmation and selection biases, not valuable information or data. Also, apparently Vanderbilt and the University of Tennessee are just as bad ...

Actually, wait, I found a useful piece of information in this I already knew. College students send the most tweets and the activity centers are university towns.

_________________

So your argument is that there is too many unknown factors and biases in this data collection that it can't be used to infer anything at all.

That's possible.

I would argue that the study does show us rather reliably that, based on the percentage of tweets that can be verified as originating from either side of the USA , eastern tweets are more likely to contain the words "nigger" or "fag" (or whatever they're counting) than the percentage from the western half of the USA.

Beyond that, tells us nothing. I entirely disagree with the premise that a person can reliably tell how the word is being used. I entirely disagree with the premise that it shows a predeliction for racism in some way. It only tells us that among twitter-users who allow geolocation, eastern tweets are more likely to contain certain prejudicial terms.

I'm not sure I care about that information, as I believe we need a prejudicial epithet for rabid twitter users in general.

_________________
Well Ali Baba had them forty thieves, Scheherezade had a thousand tales
But master you in luck 'cause up your sleeves you got a brand of magic never fails...
...Mister Aladdin, sir, What will your pleasure be?
Let me take your order, Jot it down -You ain't never had a friend like me
█ ♣ █

Of course you would. You're a woman.

_________________
Buckle your pants or they might fall down.

That's just off-topic hate. You need to hate according to the criteria.

_________________

She's also Canadian. And gay.

_________________
Buckle your pants or they might fall down.

With the addition of the more specific caveat in red, I agree. That is all that this map shows.


Now, the addition could, by some, be dismissed as "semantics." It should not. This is because whatever method Twitter uses to geotag the origination location could be falsified, erroneous, or incorrect for some other reason. Without data on the accuracy of the geotag, it's important to note that no external validation was done on that particular data point. This is important because it shows that even this minor conclusion we're drawing could itself be flawed (although admittedly not likely).


Edit: which goes to one of my original thread comments: what this thread (and the OP map) does, is show a great case example of why studies and statistics should be examined with great vigor when it comes to methodology. The map originally was claimed to show how much "more hateful" one portion of the US is than others, when in reality all it can show is unverified language patterns without causal ties to anything else.

_________________
“The duty of a patriot is to protect his country from its government.” - Thomas Paine

For shits and giggles, here's another map (spoilered for size), compiled by a different group using a different methodology:

It makes you wonder what, for instance, the Chinese respondents meant when answering that query. Mostly, I would suspect they meant other Asian ethnicities, but from another perspective, that's just yellow on yellow hate, so does that really count?

_________________

Link?

_________________
Well Ali Baba had them forty thieves, Scheherezade had a thousand tales
But master you in luck 'cause up your sleeves you got a brand of magic never fails...
...Mister Aladdin, sir, What will your pleasure be?
Let me take your order, Jot it down -You ain't never had a friend like me
█ ♣ █