Researchers "Geocoding the Blogosphere" Portend Stronger, More Authentic Detection and Analysis of Social Trends

Blog content analysis permits 'weather reports' on socio-politics of cities, regions

Release Date: March 15, 2005 This content is archived.

Print

BUFFALO, N.Y. -- The popularity of personal weblogs, or "blogs," has increased rapidly in the United States, particularly since 9/11, and they have demonstrated their usefulness in assessing social and political trends, aiding in world-wide crisis communication and in garnering political support and funding.

Despite the vast number of Internet blogs (4 million today, compared to almost none in 1998) relatively little attention has been paid to their microcontent until now.

Researchers in the University at Buffalo's School of Informatics have undertaken a long-term research project to study how information from blogs produced in specific American urban areas reflects the political agendas, opinions, attitudes and cultural idiosyncrasies of the general population of those places.

"It is our contention that the totality of content across millions of weblogs vividly and objectively depicts the social landscape and ideology at certain points of time and space," says Alexander Halavais, Ph.D., assistant professor of communication in the UB School of Informatics.

Halavais, who is conducting the study with Jia Lin, a doctoral candidate in the school, predicts that the study results and the processes used to devise them will be of use to public-opinion and market researchers whose research cannot be fulfilled without knowing the geographic location of their subjects.

As a first step, Halavais and Lin indexed 300,000 existing weblogs to their geographic locations, a difficult task they accomplished by reviewing the bloggers' ICBM meta tags; city locations inferred from local weather information linked from the blogs' index pages; the blogger profiles at hosted logs; profiles on "Blogchalk," a major commercial index of weblogs; registrant addresses from "Whois" data, and from other keywords on the blogs' index pages.

They found that:

* American blog distribution correlates positively with the distribution of the U.S. population, with most bloggers heavily concentrated in large cities in coastal areas and their surrounding suburbs (New York; Boston; Los Angeles; Chicago; San Francisco; Washington, D.C., and Philadelphia)

* New-technology and economic centers and clusters have formed large groups of bloggers. Among them are the San Francisco Bay area; Austin; Houston; Atlanta; Orange County, Calif.; the region east of Phoenix (Mesa, Chandler and Tempe); Las Vegas, and Portland

* Suburbs and regions surrounding big cities such as Detroit; Washington, D.C.; San Francisco; Boston; Phoenix; Los Angeles; Dallas, and Seattle have blogger groups comparable in size to those of their center cities.

* There are very few blog concentrations in the inland U.S., particularly the Midwest.

They also concluded that the densest concentration of bloggers is found in areas traditionally associated with "culture elites" and high socioeconomic status:

* Average household incomes in most of the 26 three-digit zip code units in which the majority of blog-clusters are found are, at around $59,000, higher than the national average and in six of those 26 areas, average household income is more than $100,000.

* Some concentrations of bloggers were found in cities where household incomes are relatively low, but those were either college towns like Berkeley, Calif.; Bloomington, Ind.; Madison, Wis., and Tallahassee, Fla., or cities like Austin, Texas; Orlando, Fla.; Atlanta and Worcester, Mass., in which 18-34-year-olds make up more than 25 percent of the population.

When it comes to the microcontent of blogs, Halavais says, it used to be anybody's guess what was being discussed among these hundreds of thousands of bloggers. Not anymore.

"Existing methods of content analysis permit the codification of large amounts of blog text with its rich, unsolicited source material and easy accessibility," he says, "and such codified, easily accessed material is a valuable resource for both macro and micro social-science research."

Where such texts have been analyzed, he says, "The resulting data have been recognized for their strength and authenticity not only by researchers, but by the media and the general public."

To support their hypothesis that weblogs accurately depict the social structure, attitudes and beliefs of non-bloggers in a particular region, the Halavais-Lin study not only will "geocode" bloggers, mapping their location using zip codes and other methods, but will look at whether and how the spatial community of bloggers relates to their geographical proximity.

"Bloggers generally don't talk to one another in a literal sense," Lin says, "but through 'pinging,' commenting, tracking and hyperlinking they form a virtual community online. So we will test part of our hypothesis by examining hyperlinks among American blogs."

"Then, using content analysis," says Halavais, "we will plot the actual blog text onto the distribution map we have created, making it possible to assess social trends in specific areas of the country -- to offer a 'social weather report' for a city or region."

Halavais and Lin acknowledge that bloggers don't fully represent everyone's opinion, pointing out that only two percent of the population produces blogs, and certainly not everyone reads them.

"Just a few years ago blogs were considered the work of a cult of computer geeks and their points of view were seen as out-of-the-mainstream," says Lin. "This has changed, however. Today it is hard to ignore the public opinions reflected in the blogosphere. Major bloggers are understood to represent much more than their personal point of view."

Halavais notes, for example, that the blogger Joi Ito is a mover and shaker in Washington, and was a speaker at the January World Economic Forum in Davos, Switzerland. 

"His primary work is as an entrepreneur and venture capitalist," he says, "but his blog is very, very popular and he funds other blogs, as well.

"Blogs like Ito's , Andrew Sullivan's 'The Daily Dish' at and Boing Boing -- which had more than a million unique visitors in December -- have many hundreds of thousands of regular readers and are of great interest to the mainstream media and to political trackers," he says.

"We cannot ignore the blog," Lin says. "It is a rapidly emerging political and cultural entity whose importance is likely to increase. It is our contention that blogs not only tell us about those who write them, but quite a bit about particular urban areas in which we find them."

"We know that bloggers are not representative of Americans in general in certain respects," Halavais says.

"They tend to be younger, more urban, more educated, more technologically adept. They're also early adopters and more willing to speak publicly about certain issues than other Americans, most of whom do not blog or even read blogs," he adds.

"Despite this," he says, "we suspect that bloggers are likely to be opinion leaders in their local communities, and that they indicate the opinions of large numbers of Americans on a range of issues. The demographics of bloggers may not exactly match those of their communities, but we wouldn't be interested in them if they held unrepresentative opinions.

"This needs to be proven empirically, of course, but it is assumption that drives this work."

The University at Buffalo is a premier research-intensive public university, the largest and most comprehensive campus in the State University of New York.

Media Contact Information

Patricia Donovan has retired from University Communications. To contact UB's media relations staff, call 716-645-6969 or visit our list of current university media contacts. Sorry for the inconvenience.