Meaningful Media Now: Hacking Election Coverage
In just a few days the United States will vote for its next president. At this point in the election cycle, most of the public has seen a vast array of news coverage about the race, which has covered the candidates from many angles. In this installment of Meaningful Media Now, we look at how some artists and programmers are using technology to analyze and remix election coverage. These emerging methods allow viewers to see familiar televised events in a new light, and may even have the potential for influence.VideogrepHarmony Institute was recently visited by artist/hacker Sam Lavigne who demoed, among other projects, his videogrep library. Videogrep searches through dialogue for a word or phrase and then creates a new, supercut video containing all instances of that word or phrase being uttered. Done correctly, the supercut is a way of highlighting repetitive elements or key insights. By hand, it is a long and tedious process. To the viewer, this illuminates frequently used words that they may have not noticed otherwise, and highlights a speaker’s unique way of communicating. For example, in the following video, Donald Trump saying his own name might tell you something about how Trump sees and talks about himself…or it might be a waste of four minutes.DuplitronDuplitron, created by the The Internet Archive, takes a piece of video, compares it to known existing media, and identifies repeated content. The Internet Archive and Annenberg Public Policy Center have been using Duplitron to identify when clips from one of the three presidential debates are replayed in the 26 hours immediately following the debate by various cable news programs. Ultimately, the difference in the chosen clips can make a difference in how viewers perceive who won the debate.
For example, the graph above shows the number of times one of three cable news networks, (Fox News, MSNBC and CNN) repeated a clip for each second of the final presidential debate. The largest peak represents Donald Trump refusing to say whether he’d accept the results of the election. That moment, as well as the debate surrounding it, was well covered by all three networks. In contrast, another peak is the moment Trump called rival Hillary Clinton “such a nasty woman.” That clip was frequently replayed but the surrounding moments were not. Another small peak representing an exchange about the Clinton Foundation was covered much more by Fox News in comparison to the other channels. Frequency, in addition to inclusion of surrounding moments, can change the viewers’ perception of what happened.Debate in (E)motionA more experimental project is Debate in (E)motion, developed by four students at Columbia University’s Data Science Institute. They captured a screen grab of the debate video every five seconds and submitted it to Microsoft’s Emotion API. The API recognizes faces in images and grades them on how much they express emotions, such as anger, contempt, disgust, fear, happiness, neutrality, sadness and surprise. The students stated that they hoped the experiment could provide an unbiased account of what transpired during the debate, and serve as an antidote to partisan political pundits. The results claim that Hillary Clinton showed more happiness during two debates than Donald Trump, who displayed more sadness, contempt and anger. However, the tool is limited in several ways. The screenshots taken may not provide an accurate visual representation of the moment. Because it only looks at faces, other cues such as word choice, tone of voice and gestures are missing.TransProseHannah Davis, a programmer, artist and musician also looked at the emotions displayed by the candidates during the debates and used them to create music. Davis created a tool called TransProse which takes a body of text and finds different emotions throughout it and then programmatically creates music with a similar emotional tone. In theory, because this analysis is done entirely by computer, it should provide an unbiased result. She also found Hillary Clinton’s emotions to be more upbeat and Trump’s more ominous. Her more “positive” tone and use of “active” words translated into music at a higher octave and faster tempo.These projects demonstrate how technology, used innovatively and creatively, can act as a form of analysis, piece of art, information piece, and social criticism all in one. Manipulation of media can be both insightful and blinding. Leading up to the election, enjoy the interesting pieces these programmers, researchers and artists are creating, but also, as with any piece of media, view them with skepticism. Many of these methods are still susceptible to inaccuracies.