Data Art

Week 5, 6, & 7: Exploring Political Messaging


Words matter.  So we are taught to be thoughtful about the words we use.  Yet, in todays political climate, even word choice has become a tool to polarize.  Every day we consume content that contains words and phrases that deepen our divide.  

History has proven that political change and progress happens when people come together for a common cause.  But how can we “find common ground when even our language is polarized” (1).  In a series called “Undividing America,” Business Insider is trying to highlight the aspects of our culture that deepen our divides in order to learn how to “undivide.”  One central cause of this division lays in the words we use to describe political issues.  We need to amplify what unites us and get beyond what divides us.  Of course, we are sometimes going to have differing of opinions, and we should, but ideas should persuade opinion, not careful word-smithery.  

From cable news to the house floor, it has become the norm to use carefully constructed phrases to make a point or to discuss a political issue.  “Behind the scenes, experts from both parties are busy devising and testing new ways to frame hot-button issues designed to pull voters further into their camps” (1).  Rachel Maddow might talk about a need for gun safety while Tucker Carlson talks about gun control.  We have come accustom to Democrats discussing undocumented immigrants, gun safety, estate tax, jihadism, and consumer protection while Republicans use the words illegal aliens, gun control, death tax, radical islamic terrorism, and corporate regulations to talk about the same issues, respectively (1). 

Framing issues in this way is a relatively new approach in politics.  In this study, researchers asked participants to read speeches that were given in Congress from 1872 to 2016.  Participants were asked to guess if the speech was given by a Democrat or Republican.  Before the year 1994, individuals were able to guess the correct party a little over 50% of the time.  However, after 1994, the guesses were far more accurate.  By 2010, 73% of the guesses were correct (3).  In this article, Mark Abadi traces this 1994 shift back to Newt Gingrich’s delivery of the “Contract with America” which is the first time some of the politicalized terms we are familiar with today appeared.  Since then, Frank Luntz and George Lakoff have become the experts in messaging for the right and left, respectively, and have built modern political strategies around these frameworks. 

In some cases, when we remove polarized framing from political issues, we can actually agree.  Even for a seemingly unshakable issue like climate change, finding common ground is possible.  The image below, taken from this article, shows how when we change our language we can agree on issues related to climate change… (5)

image taken from    here

image taken from here

As soon as you remove the word climate change, which has now become such a polarizing word, and start talking about reducing pollution and finding new energy solutions, even some climate deniers get on board (5).  This was proven when Republican Trammell S. Crow organized Earth Day Texas which aimed to show that despite the political implications of the words “climate change,” environmentalism has always been a conservative idea.  Earth Day Texas brought out conservatives who care about the environment and highlighted that there is a place for both the left and the right to collaborate on their environmentalism (5).  This other graphic by the same article highlights the potential for common ground…

This comparison shows that while we may disagree about who is causing climate change, 71% of Americans do believe that our environment is being threaten.  Perhaps if we change the language we use to talk about the climate, we can work to fix the problems we all agree on and avoid getting hung up on where we differ.

In response to this research, I have developed a series of conceptual pieces that highlight the issue. The goals of each piece is outlined below.  


For this project I would like to create three pieces that come together to educate viewers on political messaging in the hopes to find common ground between the left and the right.  It is important to understand the issue and how it has become such a norm in our politics before we can take action to change it.  Hopefully, this project can bring some of this insight.

The first element of the project will introduce the history of political messaging and how it has changed throughout history.  This station will focus on the changes in language that happened before and after 1994.  Figure 3 shows the basic setup for the history section of the piece.

Figure 3

Figure 3

 The wall will represent a timeline from 1872 to 2019.  In the middle will show the pivotal year in political messaging, 1994.  Two screens will be places on either side of 1994 on the timeline and a iPad will be placed for users to interact with the exhibit.  The users will be able to interact in three ways.  The first interaction is represented in Figure 4.

Figure 4

Figure 4

Here users can select from a list of speeches given on the house floor by either a Democrat or a Republican.  Each display will be trained using a machine learning model on speeches from their respective time frame.  When a speech is selected, the screens will “guess” what party the author of the speech was a part of.  In the second interaction (Figure 5), users can select from a list of polarizing political phrases.

Figure 5

Figure 5

When they select a phrase, each screen will display the number of times that phrase was said in the speeches that they were trained on.  In the final interaction for this section (Figure 6), users will be able to choose from a list of politicized words and the respective screens will display the word that is most commonly associated with the selected word based on its respective training data.

Figure 6

Figure 6

 In the next section of the project I would like to create an installation where users can become exposed to the current state of political messaging in our media.  Figure 7 shows the idea of this display.

Figure 7

Figure 7

The wall will be divided in half based on the political divide.  On the left, a screen will show a live stream of MSNBC.  On the right, a screen will show a live stream of Fox News.  Both streams will be muted except for when someone on the stream says one of the previously determined polarizing phrases.  At that same moment when the stream is unmuted and the phrase is heard, a count of how many times each news network has said that phrase in the past 24 hours will be projected on the wall.  This will hopefully make viewers more aware of how common and frequent this language is used today.  In the last section of the project, I would like to create an interaction where people can start to think about where we can find common ground by showing how we can de-polarize some of our language.  Figure 8 shows this interaction. 

Users will be able to type a polarizing term into the computer and a program would spit back ideas of ways to talk about this topic without using framed language. 

All of these elements of the project are initial ideas.  Once I begin to work with the data for these pieces, I imagine that the outputs may change.  No matter the ultimate output, I would like to have these three distinct sections that highlight the history, current and possible state of the issues of political messaging and framing. 


As an initial exploration of this project I created a small website shown in Figure 9.  This website uses the transcripts from the Rachel Maddow Show and the Tucker Carlson show on the night that the impeachment inquiry into President Trump was announced by Nancy Pelosi.  The website allows you to search any key word and see how many times in these respective shows that word was said.  Using this example, you can start to see how a larger project on political messaging could highlight a lot of the divide we are seeing in todays media and politics.

Eva Philips