Innovations fueled with Sentiment analytics - Revolutionizing Big Data !
It may be an attitude towards something, a feeling, or anemotion. We all have sentiments on a daily basis and we express them through various media, such astalking, SMS-ing, chatting, painting, dancing, etc... In the past few years, we have expressed our sentiments through popular social media such as Facebook, Google+, Twitter, blogging, etc.. and by using hash-tags.
People broadcast their emotions on social networking sites to let their friends and the whole world know what they feel about certain things. The study of these sentiments is a vast field with some great complexities and I won’t attempt to cover all the points in this blog post. However, I will try to convey something relevant about the analytics done on sentiment.
For the purposes of this blog post, I will be using Twitter as the medium to describe my sentimental analytics. As studies have shown, on an average, 140,000 tweets are broadcast every second. You can imagine the enormous number of tweets that are produced each day, and more so during the span of a year. As you can see, a staggering amount of data is generated everyday which is can be analyzed for useful information. Along with information that would be useful for marketing and sales, this enormous wealth ofdata contains peoples’ views, emotions, opinions.
In any given sector or field, (ex. Sports, Politics, Technology, Entertainment, Business, Breaking News, Music, Brands, Healthcare, etc?) people post their sentiments on them on daily basis.
Now the question is, “How can we analyze these emotions”. Also, once we have analyzed and understand the sentiments, how can we utilize them to our benefit. Special algorithms need to be designed in such a manner that positive and the negative comments can be identified. For example I put "This restaurant has a great ambiance" -Definitely Positive. "Never go to this restaurant because their service is very poor! " -Definitely Negative. Here you can easily differntiate between positives & the negatives. It is not always easy to differentiate between a fact and an opinion. When some sentences are tweeted with slang or sarcasm, it can be difficult for the algorithm to determine the true emotion being expressed. For instance "The food was so great that I would rather save it for later." - Sarcastic and negative. Another example, "I had one hell of a meal at this restaurant or This chick looks was f*** hot while dancing" ? These statements contains both slang and negative words but actually expresses a highly positive sentiment.
Its a difficult task to differentiate between a negative and positive emotion when slang sarcasm are used together. It’s the job of the Analytics team to give the right analysis on the views people are posting on social media. The US Presidential election & India's General Elections are classic examples of how social sentiments played a major role indetermining the mood of the Nation.
Here we are using more than 1.67 billion Twitter data points to determinesentiment/trend for any popular event. Let’s take FIFA World Cup 2014 for an instance. This massive event is all set to kick off very soon. I would like to track the trend of this event, so I used "FifaWorldCup" as my keyword and placed that keyword on my "Social Sentiment Analyzer".
You can see Analytics chart. Most of the dotted regions are based in Positive section of the Chart.
Also, one can get the further details via heat maps, Tag Cloud, TimeLine, Map, Breakdown, Comparison, Tweets, etc? There is a infinite number of opportunities to generate analytics and reports from the available Sentimental Analytics.
I believe that by now, you would have gotten a general sense of a real-time Sentimental Analytics. You can see that thepossibilities are endless and the potential for insight is amazing. I feel an out of the box approach can really revolutionize the Big data Industry. We can realize thepower of sentiments and thehuge scope which this industry has, rather than thinking of Big Data as a Pandora's Box.