My Thoughts on Social Sentiment Analysis
I’ve been watching a few folks bantering back and forth about “sentiment analysis” and the like. I have to admit that it has been very hard to keep “scientific girl” under wraps and not question the rationality of the conversation.
With a background in chemistry, math, and information technology; I’ve mostly dealt with data that are absolute and measurable. The word “data” has Latin origins meaning “something given.” I have a difficult time acknowledging the measurement of human emotion by analyzing linguistics.
There are commercial software programs ($$$$$) that use various algorithms and theories in an attempt to analyze the humanity behind the written word. Most that I’ve looked at require the input of taxonomies and or ontologies, speech lexicons, and other configurations prior to analyzing word streams. Despite the extensive pre-work, I fail to see how any of these could quantitatively evaluate emotion, sarcasm, slang, or other human communication traits.
There are open source programs like UIMA and LingPipe that can be used to build frameworks for linguistic analysis. Jumping into either requires much more headspace that I care to lease out at the moment!
If you continue to watch this segment closely you will soon see a rapid emergence of companies clamoring for market share. Some offer human insights, others do not. In my opinion the best article regarding social sentiment analysis was published by Jason Falls (@JasonFalls) entitled “Why You Shouldn’t Trust Automated Sentiment Scoring.”
Don’t get me wrong: I’m not saying the people behind automated scoring aren’t working hard or helping us accomplish a difficult task easier. I am saying, however, that we need to be clear that letting a machine supply us with this particular piece of marketing intelligence is flawed. It’s not that we shouldn’t do it, but that if we do, we must understand the limitations and prioritize the intelligence accordingly.
That pretty much sums it all up! Perhaps there will be a day sometime soon when computers can truly be emotive and adaptive like HAL. But until then, we all need that human intuition and touch and I’m not sure that can be shrink-wrapped and sold.
What then is the answer for the small business owner? At this point in time if your brand (or company) isn’t yet widely recognized, it would be difficult to justify the expense of social sentiment analysis using some of the currently available commercial tools. If you’re just curious, you may want to take a look at some of these no-cost alternatives.







