Live from SXSW: Data Visualization For Social Good

March 19, 2015       Joleen Ong      

Dr. John Snow mapped the cases of cholera in London’s Soho neighborhood in 1854, thereby tracing the source of the outbreak, and inspiring reform in water and waste management in London and beyond. Deemed the “father of modern epidemiology,” Snow saved countless lives through his data visualization.

More than a century and a half later, data visualization’s surging popularity can be seen in the countless free and paid tools and services to create your own visuals, new academic programs, and high number of jobs in this particular field. In the U.S. alone, some labor analysts believe that people are needed to fill an estimated 1.5 million data interpretation–related jobs.

“The popularity has to do with the incredible access we have to data, and the spirit of data transfer,” explained Aaron Hill, assistant professor and director of the Data Visualization Masters program at Parsons The New School for Design in New York. “Before, you needed to use the Freedom of Information Act to get data, and now you can get anything.”

During the final day of SXSW Interactive, Hill led the workshop, Data Visualization for Social Good, with a packed group of attendees. Hill, who played a key role in developing the curriculum for the new Data Visualization Masters program at Parsons, shared the process behind creating data visualizations for evidence-based decision making, and guided attendees to prototype their own using the City of Austin’s open data repository.

Hill shared, “a common question that people ask when getting started with data visualization is, ‘what software should I use?’ but they really should be asking, ‘how should the data map to the visual representation?’”

Perhaps the most surprising question that Hill encouraged people to ask at the onset of any data visualization project is: “What’s the human factor?”

Audiences looked at two data sets — bicycle rack requests and Austin’s public art collection, and were tasked with deciding how to use and interpret this data to tell a story about equity in Austin.

In response to reviewing Austin’s public art collection, one attendee explained her perspective on how to view the data through an equity lens. “I would cross reference the data with real estate values and surrounding quality of life data, and then overlay this information with other social data, such as the profitability of local business, taking into consideration factors like the broken windows theory,” the person said.

Hill closed the session by sharing his three principles for visual representation.

Principle 1: Graph Grammar

Think about how points, lines, and shapes can best represent the data. Hill cited the “graph grammar” work of statistician, Hadley Wickham, and emphasized the importance of consistency in the how data is described.

Principle 2: Layering and Separation Techniques

It’s important to emphasize the layers that directly map to data. Hill shared an image of an airport diagram map and urged attendees to think about the data points that in reality actually matter to pilots: location of the runway, and areas on the runway that have historically been prone to accidents. One version of the map that included water was

Principle 3: Graphical Perception

It’s critical for data to be presented in a way that supports accurate interpretations. Hill shared an image that compared the same data in a pie chart versus a bar chart, and explained how data presented in circle can be hard to convey precise information, especially without the baseline that a bar graph provides.

The complete set of slides from the presentation can be viewed through the link: tinyurl.com/sxgood.

As explained by Hill in an earlier interview by the New School, “Data visualization democratizes our understanding of data and research. Visualization can convey results of an analysis more elegantly, with a primary goal of being comprehensible and revealing.”