Luke DuBois: Portraits in Data

Data as the 21st-century color palette

At the most recent GovLab Ideas Lunch, R. Luke DuBois shared his work and perception on data and arts in the GovLab offices. DuBois holds a doctorate in music composition from Columbia University, and has lectured and taught worldwide on interactive sound and video performance. He has collaborated on interactive performance, installation, and music production work with many artists and organizations and was the director of the Princeton Laptop Orchestra for its 2007 season. He is currently the director of the Brooklyn Experimental Media Center at the NYU Tandon School of Engineering, and is on the Board of Directors of the ISSUE Project Room.

DuBois views his work as a journey in redefining the concept of human portraits using data. In the below we delve into some of the examples Dubois shared during the talk, as well as the key takeaways from his diversity of projects.

Luke DuBois: Digital Era Artist

Luke DuBois: Digital Era Artist


Key Takeaway #1: The medium chosen to visualize data adds another important layer to the message received. In a world flooded with endless stream of information, finding ways to reimagine commonly viewed and understood objects (like maps), can create new insights and push the audience to rethink the way they interact with information.

Project: Presidential Eye Test

“The Medium is the Message” phrase probably describes best DuBois’s “Hindsight is Always 20/20” project. DuBois used the University of California’s American Presidency Project data of 220 years of President’s Annual Messages to Congress (better known as “the State of the Union Address”) and represented the most common words used by each of the Presidents in an eye test pattern format (known as “The Snellen Chart of optical acuity”). The result was 42 eye-tests charts representing a fascinating angle of presidential agendas, concerns and anxieties.

President Reagan's most common words in his State of the Union Address represented as eye-test charts

President Reagan’s most common words in his State of the Union Address represented as eye-test charts


The project is an example of how DuBois uses his unique ability to take huge datasets and leverage his computer science skill-set to extract insightful new points of view on culture, society and politics.

Project: U.S dating Sentiment Analysis

Another example of the ways the chosen medium plays a game-changing role is DuBois’s “A More Perfect Union” project. This is probably his most ambitious and inclusive project in terms of data collection and analysis. As a tribute to the national U.S government census (that occurs every 10 years) DuBois broadened the conceptual lenses of the census and added layers relating the emotional and mental state of the American people. To achieve this ambitious goal DuBois used more than 19 million records of 21 different dating websites (such as and JDate). DuBois used the way people describe themselves to make two types of visualized analytics: The first was creating sentiment-based heat maps of the United States using the different adjectives scales people described themselves according (e.g., shy, sarcastic or sexy).

pasted image 0-2

DuBois’s version of national census: this heat map represents how people describe themselves in dating profiles

The second piece of the project involved a road atlas of the United States, with the names of cities, towns, and neighborhoods replaced with the words people use to describe themselves and those the type of companion for whom they are looking. Each word appears in the place it’s used more frequently than anywhere else in the country.

These two parts of the project contribute to two common psychology conversations. The heat map provides an interesting angle to the ongoing “environmental vs. genetic” or “nature vs. nurture” discussion around what shapes humans personalities. We can clearly see patterns of the ways people describe themselves as a function of their whereabouts. The road atlas, on the other hand, demonstrates how strongly embedded the abstract image of some locations can be – seeming to shape the way people feel about and accordingly describe themselves (evidenced by, for example, New York City’s label of “Now,” San Francisco’s “Gay,” Seattle’s “Heartbroken”).

pasted image 0-3

DuBois’s version of California map where San Francisco is replaced with Gay, Sacramento with Afternoon, Berkeley with Young and San Jose with Architecture

Key Takeaway #2: Data visualizations can add an emotional layer to raw information, pushing the audience to react in a different way than they would when simply looking at chart with numbers. Visualizations can make fluctuations in the stock market more understandable to the average citizen, but also, for example, elicit more of a reaction to child mortality rates in a given country than a simple collection of statistics might.

Project: Hard Data, The Iraq String Quartet and Taking a Bullet for the City

During the talk, DuBois expressed his interest in shifting the definition of the term “Hard Data.” He suggested a new interpretation to the phrase – which usually describes quantitative data (in the form of numbers or graphs) – instead using it to refer to the hard emotional experience the person who interacts with the data would feel.

The “Iraq String Quartet” project is a one example of this new interpretation of Hard Data. The idea behind the work was to create an open-source “information score” of the U.S.-led war in Iraq, taking various statistics from both Iraq and the United States over the course of the war and placing them on a timeline. This timeline was then published in a format that artists and composers could use to make new works. During 2009, DuBois created a string quartet version focusing on the casualty statistics of Americans and Iraqis involved or caught up in the conflict, and “sonifies” the data stream in six movements, compressing the six years of war into twenty-five minutes of musical time. While listening to the music the tone changes represent the stream of data of the casualties among men, women, children, soldiers, the number of refugees and missing. The Iraq war, being very distant from American public (physically and, increasingly as the timeline of the engagement carried on, mentally and emotionally), represented for DuBois a conflict in which most Americans have more data than knowledge.

YouTube Preview Image

DuBois is taking this notion one step further in his latest art project named “Take a Bullet for This City.” This project uses a simple computer-driven mechanism to pull the trigger of a gun loaded with blanks in response to a shooting in the city of New Orleans, ejecting a spent cartridge into a vitrine that accumulates empty bullets. The noise and flash of the gun provides an alarm that is itself meant to alarm; the vitrine resembles a wishing well, only it represents wishes taken away, not granted. This piece is Hard Data in both senses of the word: it is based on facts; facts that are, by their very nature, intended to elicit an emotional, visceral response. Positioning the gun in a quiet gallery puts the visitors in a tricky situation which dramatically increases the impact on the gallery visitor: on the one hand you almost want the gun to go off and see the exhibit act, on the other hand if the gun indeed goes off  something terrible has happened, probably just a few miles away from you.

DuBois’s “Take a Bullet for This City” project.

DuBois’s “Take a Bullet for This City” project.


Project: The Invisible Cloud Cupid

The last example DuBois touched on during his talk is a more optimistic one. “Data for Good” is a burgeoning movement in technology sectors, as, for example, civic tech accelerators, prizes and challenges aimed at using big data analytics to address public issues are growing in prevalence. DuBois’s “Missed Connections” project suggests a different and creative implementation of this notion, addressing Craigslist’s Missed Connections channel – a platform for those seeking to find a person with whom they had a fleeting real-life connection – like a pleasant conversation on the subway during which contact information was not exchanged. DuBois’s “Missed Connections” is a web-based project that scrapes the Craigslist feed of the same name in one of several cities. It then matches ads posted by those looking to find their missed connections, looking for pairs of listings that are likely referring to one another. When data analysis yields the conclusion that two people are likely looking for each other, an automated email is sent to both parties, encouraging them to meet. How many new married couples are there thanks to DuBois’s work? We’ll never know that as the service DeBois develops sends the information from a “no-reply” domain.

Apart from being somewhat altruistic and creative, this project represents one of the most significant missions of data scientists which is to join together the “missed connections” between datasets and potential real-world applications and insights.

Philosophy behind data and art

DuBois defines his work’s goal as an effort “to redefine what we think about what a portrait is.” He thinks Americans don’t pay enough attention to words, from which we can all learn. His artwork seeks to identify and use the best metaphors and creative devices to deliver his message based on in-depth, time-intensive research. He also describes his work as “creative advocacy”: by leveraging technology and data he is able to deliver his message in memorable, unique and innovative ways that allow the audience see a different truth behind the information. The technology and the data enables his message to be delivered in a multi-layer and multi-sensory manner as seen in the examples described above. For centuries artists have used the tools of their time to create meaning and elicit emotional responses from their audience. In this way, Luke DuBois is quickly becoming the ultimate data-driven Rembrandt of the 21st century.

No comments yet.

Leave a Reply