From the Ridiculous to the Sublime: Cracking the Code on Sentiment Analysis

Sense making: we do it all the time in our daily lives and in professional decision-taking.   We observe, we synthesize, draw upon our experience, and blend in a few data sets.  

Sense Making, eric fisher, flickr

And, voilà!   We arrive at our own conclusions – possibly significantly different than another colleague observing and analyzing from his/her own perspective.

So what’s the skinny on sentiment analysis as pertains to our online world? [a world where a social site: Facebook receives more visitors in the USA than web search giant: Google]

If you read blogger ArnoldIT’s post last month on Marco Arment’s (former lead developer at Tumblr) blog about gauging the worth of an ‘appstore’ product: “Sentiment Analysis is quick and dirty.”

Positives are posts which most often include:  awesome, worth, thanks, amazing, simple, perfect, price, everything, ever, must, iPod, before, found, store, never, recommend, done, take, always, touch

Negatives are posts which include: waste, money, crashes, tried, useless, nothing, paid, open, deleted, downloaded, didn’t, says, stupid, anything, actually, account, bought, apple, already

What other points of view are there out there?  For example, what is sentiment analysis when consumed on a mobile device?  Well, it could be an mobile app called Glow.  Glow is an iPhone app that maps user emotion, displaying group sentiment through color.  Glow provides the ability to view how people in your current surroundings feel – blue color represents a “good feeling” while red shows “not feeling good” areas on your local or global map.

Santa Monica? Humbug.  LA? Whooping it up!  And a bit more far-fetched, but also available, is Glow’s “emotional terminator vision”: when paired with Street view, Glow uses your camera as a window that lets you see how people are feeling around you by mixing reality with Glow for a sentiment heat map over streets, houses and buildings.

Glow

What about in the business context; for lines-of-business directors who are looking to derive meaning from the gigantic fire hose of social data (that we know includes our companies, products and brands)?   A representative from Radian6, a social media monitoring software vendor shared this view:  “There is no sentiment analysis which is perfect due to the limitations of our current computing capabilities.”  Although, I’d argue, Radian6, in the meantime, is doing as good a job as anyone.

So where does it leave the rest of us?  We’re still reaching for new ways to gain a better understanding of our customers.   TheSocialCustomer.com’s upcoming webinar; will do just that.  Brent Leary and a panel of speakers convene October 19^th, 2010 to discuss Sentiment Analysis for Customer Service – its challenges and new advances – to make sense of it all.