Customer service is worthless if you aren’t able to hear what your customers are saying. Enter Stage Right: Sentiment Analysis. A business can use this technology to collate millions of pages of data (say, for example, Facebook comments) and understand what they’re saying. This is important for reputation management. It can also help a business identify opportunities for growth and refinement. Businesses who once had to leaf through thousands of feedback forms and phone surveys can now run that information through a computer. It then returns to you who is happy and who is sad in a matter of minutes. This technology wouldn’t change the business world if it stopped there, though. Sentiment analysis allows you to see what these groups of people are talking about specifically and how they feel about those things. It’s pretty rad when you think about it—sentiment analysis allows for a world in which you hear every customer’s voice.
Last week we published an intriguing rundown of how healthcare professionals use sentiment analysis. From predicting coronary heart disease and asthma attacks on Twitter to tracking HIV outbreaks around the world, the usefulness of sentiment analysis in medicine is amazing. It allows local healthcare providers to gauge the wellness of their communities. Further, the data provided is always updated every day, since much of it relies on social media. There isn’t much more to say for it—sentiment analysis can potentially save lives.
When it came to democracy, Athens had the right idea (I mean they better, they invented it). Citizens were integral to the legislative process and their opinion counted. More and more we hear reports of special interest groups, sagging voter turn out and big money in politics. But why not big data? Well, the Obama campaign leveraged this repository in the 2012 campaign. In December, 2012 the MIT Technology Review explored how his team was able to pinpoint not just key demographics but key people within those demographics. Further, compiling the wealth of information available across social media alone could help our politicians bring into focus what their constituents are talking about.
The prospect of sentiment analysis in crime fighting might seem unappealing at first, but it’s been happening manually for years. Law enforcement agents have been thumbing through thousands of documents for decades in an attempt to sniff out crime as it’s happening. Now, that process can be automatic. We hear too many times the narrative of the troubled citizen posting ultimatums and threats on YouTube or Facebook or an obscure forum before turning on the public. With sentiment analysis, law enforcement agents are potentially able to identify and foil these plots before they take place. You’ll find an example of this is over on data scientist Luca Foschini’s blog. He used Lexalytics’ Salience engine in an experiment to show how sentiment analysis could have helped prevent a spree killing in California based on a YouTube manifesto by the suspected perpetuator.
This is an interesting area. The Efficient Market Hypothesis suggests that investors who are able to ascertain the sentiment of social media quickly will make better trades and maximize their returns. Stock value tends to depend on news and public sentiment. When a company gets a lot of press, its stock prices tends to fluctuate depending on the sentiment. Programs that analyze this sentiment very quickly are effectively the engine behind this process.