The Future of Sentiment Analysis

2013 saw many advancements for the application of data, from Google’s driverless cars to The Guardian’s software that automatically compiles newspaper articles. A natural first thought is that these new technologies may lead to unemployment for taxi drivers and journalists, while others may simply experience that space race feeling of discovery and wonderment. Regardless of your initial reaction, if you have followed the current technology advancements, then it’s hard to go to work without thinking about machines making your job redundant via automation.

The continued rise of data is inevitable, but the value of data is still determined by the quality of analysis. And amongst the many fields of analysis, there is one field where humans have dominated the machines more than any – the ability to analyze sentiment, or sentiment analysis.

Sentiment analysis has been a safe-haven for analysts mainly because its functions and nuances were too complicated for machines to challenge. The ability to understand sarcasm, hyperbole, positive feelings, or negative feelings has been difficult for machines that lack feelings. The best sentiment analysis programs in use by data miners achieve, at most, an accuracy level between 30% and 40%, while humans are capable of a 96% accuracy. Yet in spite of the limitations to sentiment engines, organizations will incorporate them into their data analysis at all costs in the areas of customer feedback, marketing, CRM, and ecommerce.

Enter IBM’s Watson API. Watson, the famous super computer that is the reigning champion of Jeopardy now has the three letters, A. P. I. after its name. The possibilities are seemingly endless and the drool spilling out of the mouths of developers everywhere is, well… seemingly endless. They are drooling because Watson has something never before available, a sentiment accuracy of 86%. Watson understands sarcasm, hyperbole, abbreviations, positive sentiment, negative sentiment, the works.

With sentiment analysis being far out of the reach of computers past, the Watson API is well equipped for disruption. Watson is already being employed by North Face to help personalize ecommerce, hospitals with answering patients questions, and writing a quiche recipe.

The 9Lenses team is committed to staying on the cutting edge of data technology and incorporating new versions of sentiment analysis into our software platform to improve our voice of the customer and voice of the employee offerings analytics. However, we also recognize that while computers have come a long way, they do not replace the human analyst but rather augment them for more impactful insights.