Friday, June 22, 2012

The Perfect Chunk, the Twitter Pulse, and Why I Feel Depressed


When Twitter came along I thought I died and had gone to heaven, linguistically speaking.  As a practicing poet who had endured countless hours of graduate school misery to compress what I was writing into poetic bits that could be approved by the higher ups, I had finally found my calling. Even better, the process was institutionalized on a global basis.  The world would be my oyster.  Here was this Twitter thing, sounding very much like the song birds outside my window, through which I could upload 140 characters in poetic chunks.  For a moment I felt very modern, writing poetry in chunks of metadata that were searchable and, even better, would last forever.  I’ve read that chunky content in this weird tablet world is the next big thing and, no, it has nothing to do with the soup.

I am deeply saddened and embarrassed when I get the occasional lament from my friends at Twitter saying: “We have missed you at Twitter.” Well, truth be told, I have missed you as well and tried my best.  Didn’t I begin my tweets, appropriately, with short takes on the Northern Cardinal, the Downy Woodpecker, and the Red-Bellied Woodpecker?  Not a Common Tern among them!  I upped the ante and tweeted about the Red-tailed Hawk who savaged our local squirrel population.  Then I provided some night music with tweets about the Barred Owl who waits patiently for the mice to appear, offering no more than a lullaby of ooh, ooh, oohs to put us to sleep.

I guess Twitter has not made a lot of money monetizing my data.  But now that I have digested a fascinating research piece from UCLA and HP Labs, I am ready to have another go.  The paper by Roja Bandari and company, entitled “The Pulse of News in Social Media: Forecasting Popularity,” argues that despite randomness in human behavior, “it is possible to predict ranges of popularity on Twitter with an overall 84% accuracy.”

The focus for this study is news articles that by definition usually have a short lifespan.  Therefore, the task is quite daunting.  The researchers’ goal was “to discover if any predictors relevant only to the content exist and if it is possible to make a reasonable forecast of the spread of an article based on content features.”  The study collected data from Feedzilla, a news feed aggregator and developed an algorithm.  Popularity for a news article on Twitter was measured as the number of times a URL is posted.  They considered four article characteristics: the news source that generates and posts the article; the category of news; the subjectivity of the language in the article; and named entities mentioned in the article.  The researchers collected 40,000 feeds during a nine-day period in August 2011.  They refer to the average tweets per link or article as the t-density score.  Their analysis showed that technology, health, and fun stuff ranked the highest in t-density.

Perhaps even more interesting is the HP Lab team finds that the “top news sources on Twitter are not necessarily the conventionally popular news agencies and various technology blogs such as Mashable and the Google blog are widely shared in social media.”  But brands still matter.  Researchers found that one of the most important predictors of popularity was the source of the article.

We know from Twitter evangelists Jimmy Lin and Gilad Mishne that its users create approximately a quarter billion tweets and more than two billion search queries a day.  The two evangelists studied churn in tweets and real time search queries.  Churn is simply a measure of changes in term rank over time.  The researcher found that there is more churn in tweets than in search queries and that must be good for the business.

The folks at gigaom.com wrote the headline: “Twitter Slowly Unfolding Its Search Ambitions.”  This would include the recently unveiled expandable tweets and hashtag-based pages.  Having learned from Google about text search, Twitter’s ambitions will include more images and video.

The other day the NYT had an article on how depressives surf the Web.  Not surprisingly, they are compulsive, switching frequently among multiple application, from games, to chat rooms to file downloads.  Of course, they also show anxiety about email.  Perhaps they have a Mother Complex.  At any rate, this is called Flow Duration Therapy.  What good is a disease without a name?

Our friends at the Missouri University of Science and Technology gave us this study. One possible solution proposed is the creation of a software program that would tell us when our Internet usage is showing a pattern that might signal symptoms of a depression.  A perfect solution!

Psychologists refer to the “availability bias” to explain why people make certain decisions.  Amos Tversky and Daniel Kahneman, Professors at Stanford University and the University of British Columbia, respectively, use this phrase to explain why people don’t always make rational choices.  Others have used this theory to explain why so many people jumped on the Facebook IPO.  It wasn’t necessarily because of the financial proposition.  As Rebecca Waber notes in her recent Harvard Business Review Blog, the critically important factor for many investors likely was "Facebook’s ubiquity and its starring role in so many people’s lives.”  Because so many of us use Facebook every day, the company likely loomed larger in the mind of investors than its balance sheet warranted.  This is the “availability bias.”

The “availability bias” might have implications for how we engage, consume, and invest in social media.

No comments:

Post a Comment