Pages

May 27, 2014

Hey Y'all! Geographies of a Colloquialism

Here at Floatingsheep, we've spent the last several years trying to demonstrate the potentials, as well as the pitfalls, of using user-generated internet content for geographic research [1]. A key focus has been how the online world of social media at times reflects, and at other times distorts, our understandings of the offline, material world. 


With all the recent hoopla around the geographies of language, we wanted to return to this topic, using a relatively straightforward example: the geography of y'all. No, not the geography of each and every one of you, the geography of the word "y'all" (see definition below). 


Rather than conducting a survey to measure the term's usage, we decided (after careful thought and rigorous debate) to do something new and use geotagged tweets [2]. Searching all of the geotagged tweets in the United States from July 2012 through March 2014 for variations of "yall" (the most commonly used y'all, as well as ya'll and yall to capture typos or alternative spellings), we found a total of 1,870,687 tweets using this folksy second-person plural pronoun, more than enough to make some definitive conclusions (or at least some maps).

Using only the absolute number of tweets with references to y'all to begin, this is clearly a geographically-specific phenomenon. While some places are extremely saturated with references (we'll get to these in just a sec!), there are 250 counties in the United States with no y'all tweets whatsoever, and approximately 60% of the country's 3,143 counties had fewer than 100 y'all tweets in the nineteen month period from which our data originates.

Still using only these absolute numbers of tweets referencing y'all, Texas, Georgia, Florida, North Carolina and California make up the Top 5 states, while the cities of Dallas, Houston, Chicago, Philadelphia and Los Angeles make up the Top 5 metro areas. And while not exactly mimicking population distribution, there is something clearly suspect about believing that folks in Los Angeles, Chicago and Philadelphia say y'all more than good old fashioned Southerners do. So, to make the map below, we instead normalized the county-level data by the total number of tweets originating in those counties during the same time period.

Geotagged Tweets Referencing Y'all, July 2012 - March 2014

On the broadest level, all suspicions and previous research on the matter is confirmed using our normalized tweet dataset: y'all is much more likely to be uttered (or tweeted) in the South than in any other part of the United States... or even the world, for that matter, as there are approximately sixteen times more references to the term in the USA than in the rest of the world combined [3]! But even still, there are some interesting anomalies worth commenting on...

Using these normalized values, we can see a new hierarchy emerge at the state level, with Louisiana, Alabama and Georgia having the highest relative number of tweets, much more in line with what one would expect. At the county level, 97 of the top 100 normalized values are located within the south (by practically any definition). The only three counties outside of this region in the top 100 are Boundary County, Idaho, Dawson County, Montana and Goshen County, Wyoming, the first two of which surprisingly rank #2 and #3 overall in these normalized rankings, led only by Talbot County, Georgia, the epicenter of y'all-related tweeting [4]. But even the South isn't homogenous when it comes to the usage of y'all, as the central Appalachian region of eastern Kentucky, West Virginia and southwest Virginia remains relatively untouched by Twitter references to y'all, despite being more-or-less surrounded by them. Indeed, Kentucky (spiritual homeland of Floatingsheep) is relatively sparse in references to y'all, despite selling these extremely expensive sweatshirts that attempt to capitalize on the state's southern charm.

Apart from some of these slight anomalies, much of this should come as no surprise to anyone who has spent much time in -- or even knows somebody from -- the South. So we thought it might be interesting to compare our own map to a handful of similar maps that have been circulating around the internet recently.

Some Other Maps of Y'all

The first map shows a stark north/south divide between the places that say "you guys" and those that say "y'all" (and, well, Pennsylvania, the western portion of which is also known for its use of "yinz"). The second map, taken from the New York Times interactive dialect quizdeveloped by Joshua Katz, largely resembles our own map, but seems to place the epicenter of y'all much further west than our own, in southeast Louisiana, bleeding over somewhat into Mississippi. 

So while there is some general agreement that Louisiana, Mississippi, Alabama, Georgia and Texas form the territorial heart of y'all, our work, along with the data from the Times' dialect survey, disputes the cut-and-dry story told by the first map. While it shows significant portions, if not all, of Missouri, Oklahoma, Arkansas, Kentucky, West Virginia and Tennessee, among others, firmly in y'all country, the dividing line appears to be both quite a bit further south, and quite a bit more squiggly [5] in nature. While some conventionally Southern states have only relatively confined pockets of references to y'all in our dataset (as well as in the Times' data), it's equally important to recognize that there are pockets of y'all densely concentrated in some more far flung areas of the country as well.

But ultimately as long as you have a group of friends worth using a second-person plural pronoun -- contracted or otherwise -- in reference to, we imagine you're doing just fine. 

Y'all come back now, y'hear?!
-----
[1] Wait, wait, wait... there are pitfalls to this?!?!?!?!
[2] This was also the most convenient data to use, since we had them lying around.
[3] The Bahamas and South Africa come in at #2 and #3 globally in references to y'all.
[4] We suspect that Talbot County, Georgia is the epicenter of exactly nothing else. Although we fully expected that someone from there will angrily correct us very shortly.
[5] That's a technical cartographic term.

May 22, 2014

Tooting Our Own Horn and Other Updates

Loyal followers of the blog have surely noticed a sharp decline in the number of blog posts in recent months, for which we are truly sorry and attempting to remedy the situation. We do, however, have some pretty good excuses for why we've been turning a blind eye to your plight.

We've been busy.

But with another academic year in the books, we hope to ramp up activity on the blog and give you all plenty of maps to keep you feeling cool throughout the summer. But first, we thought we'd fill you in on some of the happenings that have kept us away from the blogosphere this year...

In addition to spending the year in Estonia on a Fulbright, our fearless leader Matt has been busy getting promoted to full Professor at the University of Kentucky. Matt has also been named co-editor of Big Data and Society, a new journal that will surely have some Floatingsheep research within its (wholly digital) pages before too long. Ate was also recently announced as an assistant editor for the journal!

Not to be outdone by Matt's promotional prowess, Mark also a received a promotion to Associate Professor and Senior Research Fellow at the OII, and became a Research Fellow at Oxford's Green Templeton College; and Monica has decided to take up a new post as Assistant Professor in the Department of Geography at the University at Buffalo this fall.

Mark has also recently been awarded a five-year ERC grant to put together a team of three full-time researchers to study 'knowledge economies' in Sub-Saharan Africa. He's building a website for the project at the moment, but you can read more about it in the meantime here.  

The only ones of us left plodding through graduate school, both Ate and Taylor have reached ABD status in recent months, with only the dread of a dissertation ahead of them. They have, however, found the time to buy a house and move back to the Bluegrass from Massachusetts, respectively.

And while there have been too many publications and presentations to list (toot! toot!), we're also pleased to announce that Mark and Matt's paper on the geolinguistic contours of the web has been named the winner of the Ashby Prize, awarded annually to the best paper published in the journal Environment and Planning A.

We hope you'll all forgive us neglected the blog recently due to these happenings. We promise to do (at least a little) better.

May 20, 2014

The Epic Tweet Fight of Bronies and Juggalos

These days, it seems like the juggalos just can't keep themselves out of fights [1]. From being investigated by the FBI for gang activity, to being implicitly blamed by David Cross for the mass extinction of polar bears, juggalos are under attack from all angles. And then there was this. Last fall, this flier appeared in one of the Floatingsheep home bases - Lexington, Kentucky - announcing an "epic street fight" between the "gangsterous Kenwick Bronies" and the "badass North Side Juggalos"

Quick links for those who need an update on subcultures: bronies and juggalos

While no juggalos, and only a single brony, showed up to the fight, the fact that this got so much traction nationally points to what is obviously a fundamental antagonism between the two subcultures [2]. So we fired up the DOLLY machine in order to understand the outcome of the "epic tweet fight" between these two groups, and how it varies spatially...because that is what we do. To begin, we collected all tweets referencing referencing variations on each of the terms for the period from July 2012 through March 2014 [3]. 

Bronies vs. Juggalos in Geotagged Tweets, July 2012-March 2014

We find a total of 9,237 brony-related tweets in the United States, as compared to 13,663 juggalo-related tweets. At first glance this difference could be explained by the path-dependent nature of subcultural fandom; juggalos have existed for about 20 years, while bronies have only been around for a few. A more likely explanation, however, comes from the fact that despite their ability to complicate dominant societal conceptualizations of masculinity, bronies have lingering problem with gender-related discrimination, as seen in the clear linguistic differentiation for female fans of My Little Pony, who are instead called 'pegasisters'. Bronies are still, after all, 'bros' [4].

Juggalos, on the other hand, seem to have a much more forward-thinking and egalitarian culture with respect to gender roles -- embodied in the notion of 'Juggalo Family' -- as the plural form of "juggalo" can just as easily include women, the so-called juggalettes, as men. Should bronies hope to have any chance in this battle royale, they will surely need to remedy this gender imbalance; after all as the famous proto-juggalo Benjamin Franklin once remarked "We must all hang together, or assuredly we shall all hang separately". Good thing this is just Twitter and no blows have been exchanged (...yet!).

While spatial distributions of both brony and juggalo dominant counties are somewhat scattered, there does seem to be a tendency for Rust Belt areas to have more juggalo-related tweets, something that wasn't evident in our earlier analysis of the relationship between juggalos and polar bears. While the one hyperactive juggalo tweeter we mentioned in our earlier post again puts rural Pennington County, Minnesota at the top of the list for juggalo dominated counties, it is followed immediately by Salt Lake City, Phoenix, Detroit and Cleveland in the top five counties down the with the clown. Though there is no obvious clustering of bronies in the USA, Houston and San Francisco both have 650 or more tweets referencing bronies than juggalos, indicating a clear safe haven for these guys. And while a bit more subtle, there also appears to be a number of brony-dominated counties in the New York metropolitan area.

Given this lack of clear spatial patterns, we're forced to call this round of the epic tweet fight between bronies and juggalos a draw. But given their recent history, be sure that this won't be the last time that you see us comparing juggalos to some other strange countercultural group! After all, we do spend a lot of time on the intertubes...
-----
[1] Or facetious blog posts.
[2] Perhaps it's no surprise that none of the bronies and juggalos showed up to the epic street fight at the Commonwealth Stadium parking lot, as our Twitter analysis shows that both groups tend to keep a low profile in Lexington. The city has only 9 geotagged brony tweets, and 19 geotagged juggalo tweets during the target time period, though there has been a 44% increase in brony-related tweeting since this data was collected six weeks or so ago. Juggalos beware.
[3] The exact searches we used were "brony OR bronies" and "juggalo*", in order to capture the plural form of each term.
[4] Thanks to noted brony expert (and geographer) Jim Thatcher for some clarification on some of these matters. He also alerts us to the fact that the fourth season of My Little Pony: Friendship Is Magic is pretty bad. So, there's that.

May 13, 2014

Sheep, Cookies, and Material Things

Many people ask "why floating sheep?" and there are answers (see link above: why floating sheep?), but we thought it was time to finally admit the true origin of the name [1], rather than mumble something about digital data, materiality, and floating.

You see, sheep do indeed sometimes come attached to red balloons, as documented by Vermont-based geographers and sheep farmers Ethan Mitchell and Susannah McCandless. As shown, in the photographic evidence below, an apparent cultural practice of Vermont farmers is to attached red balloons to sheep suspected of demonic position by virtue of possessing six legs. Non-demonic sheep are expected to demonstrate their innocence by floating above the material plane.


Sadly for this sheep, it remained stubbornly earthbound, despite the addition of sticker to the red balloon.


But all is not lost as apparently the extra legs where due to a pair of quiet lambs, and there is absolutely nothing creepy about the silence of the lambs.


Or the curious examination of red balloons by a herd of sheep. Thanks for the photos Ethan and Susannah! 

This whole review of sheep, demons, balloons and silence naturally leads us to ponder questions about how material objects are influenced by the immaterial data on the web.  For example, this booth sponsored by Oreo at SXSW that "created an algorithm"[2] to determine what flavor of cream filling should be used in a 3D printing of Oreo cookies based on what is trending on twitter allowing one to "eat the tweet". Which brings up the obvious question of what a bieber or kardasian or twerking oreo would taste like.


This leads into bigger questions about how digital words have material influence on the natural environment, e.g., Can Tweeting be used to a plant help it grow?  This study aimed to find out if talking would encourage growth.  Other studies have used twitter feeds to create music by assigning sounds to particular words from a subset of users, or the emotional content of tweets in entire regions

Of course, none of this really explains the why behind the name Floatingsheep. Suffice it to say that it was determined through the construction of a specially designed algorithm.
------------
[1] Or simply add to the general fog and confusion about the name.
[2] Kudos, for providing such detailed specifics on the "algorithm". #irony

May 05, 2014

Artists, Bankers, Hipsters and the "Bro-ughnut" of New York: Mapping Cultural-Economic Identities on Twitter

It's long been established that cities are the centers of social media activity. And while there remains a great deal of internal differentiation in these cities, especially between those with access to the necessary technologies and those without, one way this kind of social media data can be put to use is in identifying variations in different cultural-economic identities that go beyond those classified by official statistics (e.g., race, income, occupation), and how these different groups make use of different spaces within the city [1].

These maps are based on all geotagged tweets sent in the New York City metropolitan area (as defined by the extent of the maps) between June 2012 and March 2014, collected from the DOLLY Project at the University of Kentucky. Using a series of keywords, we are able to visualize the spatial distribution of some of these cultural-economic indicators and identities as manifest in this data. But rather than just plotting points representing the locations of these tweets, which can tend to resemble complete noise or rather obvious patterns such as population density, we've cleaned and normalized the data in such a way as to minimize the potential effects of a small number of active Twitter users skewing the dataset, the Modifiable Areal Unit Problem and the potential mimicking of population density.

The first of these maps compares references to the often derisive terms 'hipster' and 'bro'. Although mortal enemies in the wild [2], we wanted to see whether domestication (via urbanization and Twitter) might cause these archetypes to adapt different grazing patterns.

There are a total of 12,319 tweets mentioning the subcultural signifier for those with tight pants, pseudo-progressive politics and pretentious taste in just about everything, which are overwhelmed in absolute numbers by the 239,412 tweets referencing the male, college-aged partying demographic. But through normalization techniques like the location quotient, we are able to more accurately compare the relative number of references to these terms.

The "Bro-ughnut" of New York: Tweets Referencing 'Hipster' and 'Bro'

The highest concentration of references to 'hipster' (shown in the darkest hexagons, which represent location quotients greater than one) are located in the gentrifying areas of Brooklyn such as Dumbo, Prospect Park and Williamsburg, as well as the SoHo/NoHo neighborhoods and the area around Columbia University in Manhattan, which reinforces the all-too-often-commented-upon relationship between hipsters and gentrification. These areas are surrounded by a much more extensive belt of tweets referencing 'bro', suggesting a clear spatial divide between these two groups within the city. The circular pattern of this difference gives rise to what we like to call the "bro-ughnut"[3] of New York City, which envelops the creamy center of hipsters (perhaps made from the leftovers from the cupcake shops populating the area) [4]. 

In addition to this spatial-cultural distinction between the much maligned hipsters and bros (or perhaps mirroring it), another longstanding conflict in New York City is between the bohemian artists and the monied class of jet-setting global financiers. Rather than using references to terms in the text of tweets, this map visualizes the relationship between tweets sent by users who self identified as 'bankers' (n=19,037) versus those sent by 'artists' (n=759,027). This difference in numbers also suggests a possible public policy initiative in which every banker is assigned 40 artists to support and nurture.

Tweets by Self-Identified Bankers and Artists in New York City

Rather than the clear circular pattern evidenced in the map of hipsters and bros, the spatial signature of tweets by bankers and artists shows a much more variegated pattern, but with some clear connection with offline geographies associated with these two groups. Tweets from bankers (again shown in the darkest hexagons) tend to be concentrated, unsurprisingly, in the financial district in lower Manhattan. In addition, they can be found in the high-end, exclusive residential areas of the Upper East and West sides of Manhattan, and more suburban locations like New Rochelle, Little Falls and Staten Island. There is also a concentration of bankers tweeting at the area airports, seen most clearly at JFK, but also at the LaGuardia and Newark airports. But given the overall higher number of artists tweeting, it is thus unsurprising that these tweets are much more widely distributed throughout the area, with significant concentrations of activity throughout Brooklyn, upper Manhattan, Queens, and the Bronx.

These two examples illustrate how using more context-appropriate methods for aggregating and visualizing geotagged social media data can provide meaningful (or at least interesting!) insight into the spatial distribution of cultural-economic identities in the city. While this kind of data isn't appropriate for all questions, it allows for an investigation of concepts, ideas and geographies that go beyond those captured by official statistics.

-------
[1] A version of this post (sadly without snarky footnotes) is forthcoming in the Cityscape journal, a publication of the U.S. Department of Housing and Urban Development.
[2] As documented by 19th century German archaeologist Heinrich von Kaesewurst-Schmackhaft in his pioneering studies of proto-hipsters and proto-bros in Mesopotamian epic of Gilgamesh. Kaesewurst-Schmackhaft's translation is as follows:
Humbaba’s, (father of all hipsters) mouth is fire;
his roar the floodwater; his breath is death.
Enlil made him guardian of the Cedar Forest before it became cool,
to frighten off the mortal bros who would venture there.

But who would venture there?
Humbaba’s mouth is fire; his roaris the floodwater; he breathes and there is death.He hears the slightest sound somewhere in the ForestAnd comes stalking, enclosed in skinny jeans, pretension and ironyEnlil made him terrifying guardian,
Whose mouth is fire, whose roar the floodwater.
[3] Credit/infamy for this term goes to Doyle Stevick.
[4] According to Google, the phrase "creamy center of hipsters" has never been used before. Another first for the sheep but we can only hope that the phrase is never used again.