Often the best way to communicate information about some distant event, issue or trend is to embed the news in a story that focuses on the experience of an individual. Human incidents get the public’s attention—audiences identify with and react emotionally to stories about people.
Yet in the development sector, often the real news that needs to be told is not the human anecdotes but the statistics that have been collected. But how can a non-technical audience understand a bunch of numbers? How can the public see not only a trend, but a pattern, discover not just scale, but relationships?
The field of data visualization is exploding in importance as new technologies and software help government agencies speak to their constituencies, multilateral organizations to their member states, NGOs to their donors, media outlets to their viewers and readers. It now takes seconds to sift through reams of information and identify elusive patterns, locate important outliers, or confirm gut instincts. The connections that can be made are only limited by the creativity and insights of those who have access to the information.
I recently was enthralled by the work of a number of students working with the one of the gurus of data visualization: Prof. Ben Shneiderman, the founder of the University of Maryland’s Human Computer Interaction Laboratory. What’s powerful about what his students are doing is not only are they working with both government and non-government data archives, but they are making sense of the raw statistics by using such tools and software as IBM’s free ManyEyes and Microsoft’s open-source NodeXL. In other words, anyone, anywhere can use these to better communicate meaning.
As Shneiderman notes: “The inherent complexity of social, political, and economic processes may finally become more understandable as information visualization tools for seeing temporal changes, relationships among variables, or surprising clusters become more widely used. This could lead to interesting discoveries that provoke livelier evidence-based discussions, which in turn are the basis for informed decision-making.”
Shneiderman’s students are immersed in what he calls Science 2.0. "Science 2.0 is about studying design of rapidly changing socio-technical systems. These studies are not replicable in a lab," he argues. "You have to study social interactions in the real world. Traditional social scientists have tried to understand these systems by data collection, but more effective Science 2.0 research involves design interventions to rapidly improve e-commerce, online communities, healthcare delivery, and disaster response.”
Using ManyEyes, for example, Shneiderman’s student Melissa Egan colorfully graphed an interactive bubble chart [above] of the 29 most common languages spoken in the United States, and her classmate Tak Yeon Lee created an dynamic treemap that traced the growth in global population since 1969 by countries and regions.
Using NodeXL, other classmates, Swetha Reddy and Miguel Rios, investigated patterns in votes in the United Nations General Assembly from 1946-2008. As part of that project, they tracked the votes on the 66 UN resolutions in 1962. [Your browser may generate a security warning, but please proceed. This class site is an open wiki that anyone can view, but only students can edit.]
In that visualization [right], the blue cluster is made up of “countries with democratic political systems”: including United States, Belgium, Britain, Canada, Denmark, France, Iceland, Italy, Luxembourg, the Netherlands and Portugal. The red cluster on the left, noted Reddy and Rios, “included countries with communist political systems…. The tightly grouped sub-cluster at the bottom is all the countries with pro-Soviet regimes in power like Poland, Bulgaria, Hungary, Romania, and Czechoslovakia. The other countries in this group are members of the non aligned group but voted with the Soviet Union. The non-aligned group included countries that did not want to be tied to either the West or the East is represented in green in the middle.” It’s interesting to observe that the most tightly clustered group is the majority of non-aligned countries.
Just consider what this kind of visualization could do for making clear the patterns of voting of political parties and the positions held by lobbying groups or other elites.
Then it’s interesting to take a look at the provocative project of students Jun-Cheng Chen and Hyoungtae Cho, who used an Amazon.com book dataset that tracked the function "who bought this item also bought.” In their visualization [left] the “liberal” books are the blue dots on the left; the “conservative” books are the red dots on the right. Note that very few lines link the right side to the left or vice versa.
What does Chen and Cho’s visualization prove? It helps to confirm the conventional—but untested—wisdom that people do not gather information from across the political spectrum. Or, as they wrote: “Book readers in one camp (liberal or conservative) seldom buy the books of the other…. This strong tendency might suggest that people prefer to accept opinions with the same political perspectives and to refuse those different from them.”
The take-away of all this for the development sector is that social media has now made data collection about human relationships accessible. Web 2.0 is enabling Science 2.0. In Ben Shneiderman’s words: such social media “interventions will enable new research directions that accelerate human social/civic participation and contributions to large-scale projects such as Wikipedia or Encyclopedia of Life and important national priorities such as healthcare, energy sustainability, disaster response and community safety.”