The field of sentiment analysis is growing quickly. Google Scholar reports nearly 70,000 articles including the words sentiment analysis published from 2012 to 2013; three and a half times as many as were published in the preceding annum. Assuredly, this growth is due in part to the wide array of purposes to which people apply sentiment analysis and text mining. Beyond the number of applications is the power of social media data; one study (Brynjolfsson, 2011) found that technology investments of 179 large publicly-traded firms that adopted data-driven decision making have output and productivity five to six percent higher than what would otherwise be expected; in an era when data provably matters for high-stakes decisions, being able to make new quantities measurable will certainly be a benefit. What would be the impact on your organization if you could improve output and productivity by a mere five to six percent? Or, what would be the social impact of a policy that was five to six percent more effective?

Before setting you off to conduct your own research, however, we again implore you to take care with your analyses, for they often come with consequences. Numbers, and especially measurements, tend to get reified in unhealthy ways such as IQ and BMI have been in the last two decades. Knowing this, it is especially important that you think hard about your measurements and deliver them with appropriate caveats. Consider carefully the population of people to which you can extrapolate; the users of social media are often young and urban. Furthermore, beware of contexts in which people have incentives to provide or promote biased opinions such as when writing about rivals. Lastly, carefully consider a new trend in paid opinion writing, wherein companies or advertisers hire tech-savvy authors to spam social media with favorable information about them. Detecting these strategic actors may in fact be an interesting research area in the field over the next decade. However, if you chose to analyze social data, the need will remain to be vigilant of its pitfalls. That said, the promise of social media data is great, if it can be leveraged with care.