“Words matter.” That’s the State Department’s explanation for President Obama’s canceled meeting with Philippines President Rodrigo Duterte after Duterte abrasively insulted Obama. We’ve probably all heard the trite observation, “Words have meaning.” Indeed, words do matter and words do have meaning, and so it follows that a lot of words have a lot of meaning and matter a great deal. That follows in the emerging art of sentiment analysis, a subset of big data analysis.
ANALYZING YOU
Sentiment analysis is about taking large samples of recorded human language over a particular period and, by way of some sophisticated algorithms, identifying pretty simple broad trends on particular topics or products. Sentiment analysis, for instance, can collect language from Twitter and draw conclusions and make prediction about the likelihood the Twitter crowd will accept or reject a new product. In “The New Breed of Data Analysis,” Data Quest contributor Suvro Banerjee explains, “Sentiment analysis has emerged as a strategic weapon for many organizations to gain customer loyalty and market share.”
The same analysis can focus more narrowly on a collection of recorded language produced by a single individual, too. Some report that sentiment analysis of individuals’ language can identify depression, anxiety, fear, and other emotions just by analyzing word choice, essentially. In fact, Data Informed contributor Bernard Marr reports that “experts predict sentiment analytics soon will move beyond a simple positive/negative scale and expand into classifying a broader range of human emotions.”
IN THE CLEARED WORLD
Given the growing trend of insider threats—or the fact insider threats we’re more routinely discovering—that sentiment analysis would make its way into the spy agency world isn’t too surprising. Nextgov’s Mohana Ravindranath reports that the National Geospatial-Intelligence Agency (NGA) is looking to sentiment analysis to root out insider threats. Ravindranath writes, “NGA intends to sole-source an award for a 1-year pilot to WT Government Services, which sells such a system called SCOUT Tool.” According to the September 2 solicitation description of the requirement, NGA intends to use SCOUT Tool in its Insider Threat Office (naturally), its Behavioral Sciences Division, and its Engineering Division. Specifically, NGA will load the program on enclave systems for each of those divisions. Enclave systems or networks are those that are separate from the larger network. Enclaves are sub-networks to which access is limited and very tightly controlled.
THIS IS A TEST
It’s only a test. As Ravindranath reports, “If NGA finds that sentiment analysis technology helps it identify insider threats, it intends to ‘conduct a full and open competition for future support’ . . . .” NGA must already have some confidence in the efficacy of sentiment analysis. “Recently, the agency began gathering information,” Ravindranath writes, “about how the technology could be used to measure how effective its public relations efforts are—including speaker programs and media engagements—in changing public perception about the agency . . . .”
Now, NGA will try out sentiment analysis to help determine internal perceptions about the agency. Or, as Marr puts it, “[A]s sentiment analytics grows in its ability to accurately recognize a wider range of feelings and shades of meaning, organizations will become more comfortable with the idea of sentiment analytics and begin using it in new and even more exciting ways.”
Yes, exciting ways.
