UPDATED 01:11 EDT / OCTOBER 21 2016

NEWS

Real-time Twitter analysis shows U.S. election is … too close to call?

With just three weeks to go until election day, the result is beyond any doubt. The polls show Hillary Clinton is heading for a landslide victory. But wait! Donald Trump has just pulled ahead and could yet take the White House. Sorry, no, false alarm. Clinton is surging ahead once more.

Almost every day a new poll pops up showing a different result, raising big question marks over their accuracy (especially that second one).

With so many conflicting polls being bandied around, in-memory database firm MemSQL Inc. reckons that social media analysis could be a much more accurate way to assess who’s really winning. It has started analyzing people’s sentiments on Twitter in real-time to try to show which of the two candidates is more popular.

MemSQL says there’s good reasons to believe that Twitter could be more accurate than the pollsters. For one thing, Twitter is one of the largest forums for people to voice their opinions about the candidates, with more than 17.1 million interactions during the first televised debate between Clinton and Trump. MemSQL also cites a January 2016 study from Pew Research Center, which states: “In January 2016, 44% of U.S. adults reported having learned about the 2016 presidential election in the past week from social media, outpacing both local and national print newspapers.”

MemSQL is leveraging Apache Kafka and machine learning algorithms to showcase what it calls a “compound sentiment score” in real-time for each candidate. It analyzes every single tweet that mentions the words “Hillary” or “Trump,” which were chosen as they’re the most commonly used descriptors for the two candidates. The results of its analysis are streamed directly into MemSQL in-memory platform before being visualized in the following dashboard.

“Positive scores indicate praise and esteem, while negative scores reflect criticism,” MemSQL says. “The sentiment scores are then graphed in real time every two seconds. Total tweets are summed every minute, and sentiment scores are measured as a rolling average of the current minute.”

You can read more about MemSQL’s methodology in this blog post.

Dead Heat?

As the screenshot above shows, the race for the White House is almost neck and neck, with the two candidates swapping places multiple times in the space of just a few minutes, although Clinton does seem to spend a little more time out in front than Trump does.

Given recent polls that put Clinton more firmly in the lead, there’s good reason to question whether Twitter analysis is accurate. After all, millions of voters don’t use Twitter at all, and millions more Twitter users aren’t entitled to vote anyway.

But MemSQL’s analysis could still be still useful, providing additional insight into what people are saying about the candidates on social media.


A message from John Furrier, co-founder of SiliconANGLE:

Your vote of support is important to us and it helps us keep the content FREE.

One click below supports our mission to provide free, deep, and relevant content.  

Join our community on YouTube

Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.

“TheCUBE is an important partner to the industry. You guys really are a part of our events and we really appreciate you coming and I know people appreciate the content you create as well” – Andy Jassy

THANK YOU