Data Innovators Vivian Peng

Published on August 15th, 2016 | by Alexander Kostura


5 Q’s for Vivian Peng, Communications Officer at Doctors without Borders

The Center for Data Innovation spoke with Vivian Peng, communications officer at the humanitarian aid organization Doctors Without Borders, also known as Médecins Sans Frontières (MSF). Peng, a visual artist and data scientist, discussed the role of data science in public health and the importance of understanding your audience when communicating data insights.

Alexander Kostura: How have advances in data science impacted advocacy work of organizations like Doctors Without Borders?

Vivian Peng: Data science is a relatively new field, and for many organizations like MSF, we’re still exploring the possibilities. MSF has always had a strong foundation and appreciation for data collection and analytics. Shortly after our organization was founded in the 1980s, MSF’s Epicentre, which focuses on research and training,  was created to monitor and collect epidemiological data to analyze where our operational programs are most needed.

I see a lot of potential for data science to impact advocacy work when it comes to pushing for changes in the current research and development (R&D) system or campaigning for access to medicines for a couple of reasons. First, the open source and collaborative spirit of data science is transformative. Whereas data and information has historically sat in silos and are very difficult to access, there’s now a momentum to open up data sources across industries. One of the biggest barriers to bringing down drug prices is the lack of publicly available data. While I’m sure it will take years before that information would be released publicly, the push now for open access to traditionally closed datasets is so important.

Another area we’re looking into is the Missing Maps project—a collaborative project in which a large and committed community of nongovernmental organizations, academic institutes, companies, and most of all individual mappers, map vulnerable areas in OpenStreetMap. Missing Maps recently launched an app called MapSwipe, which engages users to identify which areas have mappable items like rivers, roads, houses, and so on. Brian Rowe, a professor at the City University of New York’s Masters in Data Science program, and I see potential in using image processing and machine learning techniques to identify these areas. We’re currently talking with the folks at Missing Maps to see how we might set up a Kaggle competition to achieve this. Stay tuned!

Kostura: You describe yourself as both a visual artist and a data scientist. Where do these two fields overlap in your career in public health and current position with Doctors Without Borders?

Peng: I call myself a visual artist because I like several different mediums. I really enjoy illustrations and animations for storytelling, and this overlaps with data science in data visualizations. I like to play with R, which is a programming language for statistical computing, data visualization software called D3, and Adobe Illustrator to visualize the data that I’m working with. This is useful in public health because we sit on so much population data, but it’s always communicated via data tables or long PDF reports. That’s not the most engaging way to convey insights from data. I think the public health industry awakened to the possibilities of visual art and data science when Hans Rosling visualized the relationship between life expectancy and household income in a way that had never been done before with Gapminder, an international nonprofit promoting the use and understanding of statistics for international development.

Currently, I’m the Vaccines Campaign Online Communications Officer at Doctors Without Borders. My work involves building communications strategies and content for our “A Fair Shot” campaign, which asks pharmaceutical companies Pfizer and GlaxoSmithKline to reduce the price of the pneumonia vaccine. To be honest, it’s been a challenge to do data science work on this campaign specifically, simply because we don’t have access to the data we need. Most countries have to sign confidentiality contracts with the two pharmaceutical companies, so they can’t disclose how much they pay for the pneumonia vaccine. Without full price transparency, this makes it difficult to negotiate for lower prices.

Kostura: You’ve spoken previously about the challenge of evoking emotions through data visualizations. Why is this valuable? And how do you actually go about doing this?  

Peng: We all like to think we’re pretty rational beings, especially data scientists. We listen to what the data tells us, but numbers and predictions on their own won’t move people to take action. For example, I can tell you how many kids die from pneumonia each day—2,500 kids—that’s a fact. It’s an epidemic, but knowing that alone won’t inspire the necessary changes we need to make prevention more accessible, such as lowering the price of the pneumonia vaccine.

While data is objective, the impact of our insights is not. At the end of the day, our insights drive people to take action—whether it’s motivating people to buy a product, sign a petition, or make an organizational decision. I know from working at Doctors Without Borders that emotions can be a powerful tool to inspire action from our users. Photos and videos are great at evoking emotions from viewers, but data visualizations can sometimes lack that power.

I’ve been playing with this concept of evoking emotions from our viewers by mapping data to our senses—taste, touch, smell, sound, and sight—because this is how we naturally process data in our world. We rely on these senses to take in information from the outside world and process it to help us navigate our lives. It’s a natural progression, then, to think about how to visualize data to be readily received by these senses that create a visceral experience for our users.

Kostura: How important is it to consider different languages or levels of literacy when communicating data and insights to different populations?

Peng: It’s super important. We want our work to reach the populations in need, and that means speaking and communicating in their languages. At MSF, all of our communications materials are written or translated in the language of the countries we work in.

But once you have the language down, how do you address the levels and different types of literacy? It’s a little more challenging and has to be addressed on a case-by-case basis. Graphics and visualizations can go a long way. I remember when I taught about HIV/AIDS in Tanzania, reading literacy was not an issue in the classrooms, but health literacy was. Before we could even talk about what HIV is and why it matters, we had to go back to basics and touch on cell biology and the immune system. To help teach these concepts, we illustrated T-cells and HIV particles into little cartoon characters and gave them personalities, which helped our students remember how they function in our bodies.

When communicating data and insights to different populations, it’s also important to account for cultural references. For example, there was this one lesson in Tanzania where we wanted our students to do the “wave.” Even when our Tanzanian colleagues were explaining the concept in Swahili, the kids had no idea what we were talking about and didn’t understand how to engage in the activity. We thought they were shy, but then we realized that even though Tanzania borders water, our kids had never left the village before and hadn’t seen the ocean. We had to rethink that lesson on the spot. My lesson learned was, as a best practice, it’s always good to do user interviews and understand who I’m creating the content for before diving into the design.

Kostura: Can you give an example of how you’ve seen data science or data visualizations have a direct impact on improving public health outcomes?

Peng: Not yet, but I have hope. The best example I can think of at the moment is an online puzzle game called EteRNA. This game crowd sources Ribonucleic acid (RNA) molecule fold designs and uses an algorithm to optimize and analyze structures in the hopes of creating a better diagnostic test for tuberculosis (TB). TB is one of the biggest infectious disease killers, but we don’t have the right medical tools to diagnose someone with TB. A new diagnostic tool that originates from the results of this game would be a long time coming and a huge breakthrough for testing and treating people with TB.

Tags: , , , , , ,

About the Author

Alexander Kostura is the 2016 Google public policy fellow at the Center for Data Innovation. Alex is passionate about information and communications technologies as tools for inclusive economic growth, good governance, and social welfare. He has most recently conducted research in corporate data sharing for social good, specifically in international development and humanitarian response. Alex holds a B.S. in foreign service from Georgetown University and an M.A. in law and diplomacy from the Fletcher School at Tufts University.

Back to Top ↑

Show Buttons
Hide Buttons