Comments
Finding the Right Words With Text Analytics
View Comments: Newest First | Oldest First | Threaded View
Re: Content analysis on steroids
  • 8/2/2011 11:29:01 PM
NO RATINGS

True enough, aaphil,

I suppose the word "perfect" isn't quite right. However, when one thinks of a huge data sorting system capable of searching for countless pieces of "unstructured" data with little or no control over how that data was initially entered, the results Google gets are still quite impressive even if familiarity has taken away some of the wonder.

Re: Content analysis on steroids
  • 7/31/2011 4:27:14 PM
NO RATINGS

I'm sorry but seeing the words "perfect...system... and Google" in the same sentence always makes me raise an eyebrow.

Re: Content analysis on steroids
  • 7/26/2011 10:09:26 PM

Hi magneticnorth,

In a sense, it has already hit mainstream. Imagine a vast store of data in a wide variety of formats all organized by the perfect analytics system. You use it every time you search the Web with Google. In a sense, a similar tool except that data is constantly being ordered and reordered depending upon each successive query.

Content analysis on steroids
  • 7/24/2011 10:47:29 AM
NO RATINGS

Wow, this is like content analysis on steroids. If this hits mainstream, a number of graduate students will save themselves from a lot of dirty work.

I think one great application is analyzing informal organizational communication. It's easy to analyze formal communication (i.e. from top to bottom in the hierarchy), but all the action really happens as communication criss-crosses the organization.

Re: Predicting with words
  • 7/21/2011 9:17:04 PM
NO RATINGS

Thanks. Looks very interesting. I'll have to take a look. It may be worth revisting here at AllAnaytics.com

Re: Predicting with words
  • 7/21/2011 4:32:41 PM
NO RATINGS

Hi, Shawn.

I wrote about these a couple of months ago here.

Re: Predicting with words
  • 7/21/2011 4:30:37 PM
NO RATINGS

Joe,

Both sound very interesting and like additional examples I would love to have a look at. Would you have any links to more information on these two studies that you could share?

Re: Predicting with words
  • 7/21/2011 12:43:33 PM
NO RATINGS

The application for political campaigns is a matter of teaching a new dog old tricks.  Political candidates are a brand, just like Pepsi and Clorox.  Nice to see it being applied beyond marketing "traditional" products.

Even more impressive is sentiment analysis on factors beyond positive and negative.  Derwent Capital Markets, the British hedge fund that bases its investment decisions on Twitter activity, is focused not on "positivity" or "negativity," but on calmness.  The study upon which Derwent's methods are based measured six different sentiment categories, and found calmness to be the one that had predictive effect -- able to predict the stock market up to six days in advance.

It's important to note that the Tweets Derwent looks at are not stock-specific Tweets -- just general Tweets.  The fund is able to gauge future stock market performance simply on the "calmness" of the content of everyday Tweeters.

Another study from 2008 found that expressions of fear and anxiety in LiveJournal posts correlated with future stock market performance.

Re: Predicting with words
  • 7/21/2011 12:34:19 PM
NO RATINGS

Joe,

One of the most interesting aspects of the Webinar was a case study in which keywords were classified based on their positive or negative attributes as a way to identify potential negative press coverage rapidly and give the client an opportunity to prepare a rapid response. The implications for political and other attitude sampling, while not specifically related to business, could be profound.

Predicting with words
  • 7/21/2011 12:17:37 PM
NO RATINGS

This is definitely an important, fast-growing area.  Organizing and prioritizing is only the tip of the iceberg; there is also decision-making.  Sentiment-based analysis, for instance, is already being used to predict financial outcomes, such as with the stock market and box office receipts.  And in social media, certain keywords often result in greater engagement.

 



INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +