Sentiment Analysis and Data Visualization

Unleashing the power of Sentiment Analysis through Visualization

 

Let’s talk about a very specific application of visualization – Sentiment Analysis. Firstly, what is sentiment analysis? The process of computationally identifying and categorizing opinions expressed in a piece of text, especially in order to determine whether the writer’s attitude towards a particular topic, product, etc. is positive, negative, neutral or advance (emotions like anger, sad, disappointment). Today sentiment analysis is not limited only to a piece of text, it can also be applied to videos and speech samples.

Before diving deep into data visualization and sentiment analysis, I think it would be a good idea to actually comprehend the need for sentiment analysis form the point of view of a business that has customers – all of them. Customer satisfaction is the key to rocketing sales and consequently brand progress. This concept is limited not only to those business that involve direct monetary transactions, but also to industries like politics, entertainment etc. There are about 3.2 billion internet users in the world. Many of these take to different social media platforms to share their experience of a new product/existing product or service they came across. They do so in product reviews, tweets, comments etc; they share their thoughts in a textual form. Now, the companies can get great insights into their consumers’ minds if they could read every piece of text written to address their product or services. This will help them improve upon their services and also make decision making easier for the management on issues regarding entering a new market or launching a new product or even getting consumer insights into any competitor products.

But here’s the catch; let’s look at the following statistic. Every second, on average, around 6,000 tweets are tweeted on Twitter, which corresponds to over 350,000 tweets sent per minute, 500 million tweets per day and around 200 billion tweets per year according to Dsayce . If a business wants to analyse the response of people through twitter, it seems a humongous task to just filter out tweets corresponding to them, and then the cumbersome process of categorizing them follows.

This is where data visualization saves the day. Big data services have now grown to that extent where it’s just a matter of a few clicks to extract relevant data from a database consisting of trillions of records. Once we have the relevant tags, sentiment analysis can be performed to categorize them on the basis of emotions or sentiments. But it is simply not worth it to have 2 million rows worth of data and not have an easy way to get the crux of it easily.

Data visualization techniques can be used to put this data into numerous charts according to the usage needs. Sentiments about any particular product or service or any topic can be plotted on pie charts, line graphs, heat maps, demographic heat maps etc. Issues related to law enforcement can be addressed after a comprehensive sentiment analysis of the demography on the particular issue. Popularity of a product or service can be depicted for over many years on a timeline and this can be compared with other data analytics tools to get more insights – in this case a popularity timeline can be plotted against changing marketing policies on the same timeline.

So now yo can see what an important tool in business intelligence can sentiment analysis be. And only Data Visualization can unleash the power of sentiment analysis.