Next time you Tweet or post something to your Facebook status, beware! There is a good chance that piece of data is mined and run through a series of complex algorithms. Your feelings, opinions and sentiments could be interpreted by services that track human emotions, perhaps inaccurately for the purpose of determining “the bottom line” especially for companies tracking their product lines.
This NYT article says …
An emerging field known as sentiment analysis is taking shape around one of the computer world’s unexplored frontiers: translating the vagaries of human emotion into hard data.
This is more than just an interesting programming exercise. For many businesses, online opinion has turned into a kind of virtual currency that can make or break a product in the marketplace.
There are already a few services out there, like Tweetfeel and Twitrratr that track peoples opinions. However none of them 100% accurate. Interpreting a string of text and discerning human emotion conveyed in the string is a daunting task and an imperfect science to say the least. The algorithms typically look for keywords in a text string and assign them a positive or negative sentiment. The challenge however is accurately figuring out the human emotion associated with a series of words in a string. “Eating a handful of blue M&M’s made me hungry for more”. The word “hungry” is often interpreted as a negative sentiment. However in this context it’s clearly meant to be positive at least for the company making M&M’s (not necessarily for one’s waistline ). When this string is analyzed by an algorithm the data could be misrepresented and in the final analysis interpreted as “blue M&M’s are not very popular with consumers”.
While a lot of emphasis is given to algorithms and how sophisticated they are it’s hard not to ignore the fact that it’s the data that really drives the innovation. Recognizing and identifying untapped data sources is key and with the increasing use of social networks there is an amazing volume of data available. As companies like Facebook and Twitter open up their data thru API’s I think we’re going to see more and more of these Artificial Intelligence type applications and services popping up all over the place.