Understanding Sentiment Analysis: What It Is & Why It’s Used
By Kristian Bannister on January 26th 2015Read this article on our full site
What is sentiment analysis, how does it work, and why should we use it? Read our overview of the broad uses and benefits of sentiment analysis.
Sentiment analysis – otherwise known as opinion mining – is a much bandied about but often misunderstood term.
In essence, it is the process of determining the emotional tone behind a series of words, used to gain an understanding of the the attitudes, opinions and emotions expressed within an online mention.
Sentiment analysis uses
Sentiment analysis is extremely useful in social media monitoring as it allows us to gain an overview of the wider public opinion behind certain topics. Social media monitoring tools like Brandwatch Analytics make that process quicker and easier than ever before, thanks to real-time monitoring capabilities.
The applications of sentiment analysis are broad and powerful. The ability to extract insights from social data is a practice that is being widely adopted by organisations across the world.
Shifts in sentiment on social media have been shown to correlate with shifts in the stock market.
The Obama administration used sentiment analysis to gauge public opinion to policy announcements and campaign messages ahead of 2012 presidential election.
The ability to quickly understand consumer attitudes and react accordingly is something that Expedia Canada took advantage of when they noticed that there was a steady increase in negative feedback to the music used in one of their television adverts.
Sentiment analysis conducted by the brand revealed that the music played on the commercial had become incredibly irritating after multiple airings, and consumers were flocking to social media to vent their frustrations.
A couple of weeks after the advert first aired, over half of online conversation about the campaign was negative.
Rather than chalking up the advert as a failure, Expedia was able to address the negative sentiment in a playful and self-knowing way by airing a new version of the advert which featured the offending violin being smashed.
Contextual understanding and tone
But that is not to say that sentiment analysis is a perfect science at all.
The human language is complex. Teaching a machine to analyse the various grammatical nuances, cultural variations, slang and misspellings that occur in online mentions is a difficult process. Teaching a machine to understand how context can affect tone is even more difficult.
Humans are fairly intuitive when it comes to interpreting the tone of a piece of writing.
Consider the following sentence: “My flight’s been delayed. Brilliant!”
Most humans would be able to quickly interpret that the person was being sarcastic. We know that for most people having a delayed flight is not a good experience (unless there’s a free bar as recompense involved). By applying this contextual understanding to the sentence, we can easily identify the sentiment as negative.
Without contextual understanding, a machine looking at the sentence above might see the word “brilliant” and categorise it as positive.
How we do sentiment analysis at Brandwatch
Remember the scene in Terminator 2, when a young Jon Connor teaches the T-800 hip 90s phrases like “no problemo”, “eat me” and “hasta la vista, baby”?
That’s not entirely dissimilar to how a linguist expert would teach a machine how to conduct basic sentiment analysis.
As language evolves, the dictionary that machines use to comprehend sentiment will continue to expand.
With the use of social media, language is evolving faster than ever before. 140 character limits, the need to be succinct and other prevailing memes have transformed the ways we talk to each other online. This of course brings with it many challenges.
At Brandwatch, we employ a rules-based process to help our software better understand the ways context can affect sentiment.
We take all the words and phrases that imply positive or negative sentiment and apply rules that consider how context might affect the tone of the content. Carefully crafted rules help our software know the first sentence below is positive and the second is negative.
“I want a burrito so bad”
“I just had a burrito. It was so bad.”
The caveats of sentiment analysis
The above examples show how sentiment analysis has its limitations and is not to be used as a 100% accurate marker.
As with any automated process, it is prone to error and often needs a human eye to watch over it. At Brandwatch, we give users the opportunity to redefine sentiment if they believe that it has been wrongfully categorised.
Beyond reliability, it’s important to acknowledge that human’s expression doesn’t fit into just three buckets; not all sentiment can be categorised as simply as positive, negative or neutral.
Predictions for the future of sentiment analysis
While it’s difficult to speculate how a relatively immature system might evolve in the the future, there is a general assumption that sentiment analysis needs to move beyond a one-dimensional positive to negative scale.
In the same way that politics cannot always be reduced to a position on a left to right scale, there are other kinds of sentiment that cannot be placed on a simple barometer.
For the future, to truly understand and capture the broad range of emotions that humans express as written word, we need a more sophisticated multidimensional scale.
Can you measure skepticism, hope, anxiety, excitement or lack thereof? Until this happens, sentiment analysis is (literally) one-dimensional!
Organisations will certainly become more aware of the applications of sentiment analysis within their marketplace, fueling the growth of sector specific services and technology delivering sentiment specific use cases – for example, intelligence tools that aid decision-making for financial traders and analysts.
We will see a shift in perception of the reliability of sentiment analysis. Users will become more comfortable with the idea that the automatic analysis of individual text material is hard to match human performance.
The insight that can be gained from large datasets (millions of Tweets) will overshadow the concerns about reliability at a granular level (a single Tweet).
Instead, the focus will be on how to make results interpretable and actionable. In the meantime, we’ll be ensuring we are working at making sentiment analysis as accurate and easy to understand as possible.