Campaign managers who urge politicians that they have a certain state’s unwavering support can now use data analytics to detect cities and even individuals within those states who may not back the candidates.

Rick Hutley, program director of analytics at the University of the Pacific in California, leads a team of students pursuing master’s degrees in data science who found a way to further break down political leanings within states. On All Analytics’s radio broadcast titled “Analytics and the Making of a President,” Hutley said that data analysis has advanced in a way that can improve how campaign managers examine support for their candidates.

Hutley introduced the possibility of narrowing a red or blue state to a red or blue city, group, or individual.  He said campaign managers could knock on individuals’ doors and deliver specific messages, rather than knock on all the doors in a city and deliver the same message.

“A state’s a pretty big place. So when you say a state is one color or another, how much of that state is that particular color? If you think that state’s in your back pocket, just how exposed is it? If you ignore it completely, you may find that it swings against you and takes you unawares,” Hutley said. “The more granular you can get, the more insight you will have. You will know where you are truly strong and why you are truly strong in that area.”

Over the course of the primaries, Hutley and his group of students analyzed the texts of candidates’ speeches. The breadth and sophistication of vocabulary, as well as the length of speeches, have different effects on certain audiences.

He and his team analyzed textual data from Republican presidential candidate Donald Trump’s speeches and determined that the businessman’s short, blunt diction could have a positive effect on many of his listeners. Hutley said that Trump’s campaign, like any other, is interested in analyzing data to provide insight in making decisions for presenting certain policy points.

“One thing Trump does very effectively is speak with a very small vocabulary. Most candidates speak at the fifth- or sixth-grade level; he speaks at the third-grade level,” Hutley said. “He speaks in very short sound bites, almost tweet-level sound bites. The advantage of that is that smaller vocabulary is able to be understood by everybody. If you use a much larger vocabulary, that’s not so easily understood by some and is only understood by a few. Does Donald Trump truly have a smaller vocabulary than his opponents, or was it smart campaign management to deliberately use this smaller vocabulary because he will reach a broader audience?”

Hutley said that data collected and analyzed by campaign managers across the primary season falls under the larger phenomenon of big data. Although people have been predicting election outcomes for decades, big data allows for more incisive study. Data analysis can inform candidates of the areas, geographic or topical, in which they are failing.

Although he admitted that big data was not a particularly descriptive term, Hutley said it allows for an increased volume, velocity, and variety of information.

“It’s real time. It’s minute by minute, second by second,” Hutley said. “We can now tap into the public and find out what they’re thinking and saying. That’s fundamentally what’s changed.”

Read More About