Excel-ling in Storytelling
Convoluted charts, plain bars, drab plots, scattered points, and swaths of bright colors and structures. And who could ignore the endless, impassive lines of black-ink digits and decimal points that are behind it all?
In the past, I had thought that data visualization was a superfluous, but occasionally appealing visual addition to the scientific explanations that comprise a project. So long as the research question behind a project was clear, I figured the corresponding data would follow.
However, listening to Dr. Eric Monson, a data analyst of Duke Data & Visualization Services, speak about his work at a Huang Science Communication seminar helped me to focus my thoughts on the ways that we could simplify and better illustrate complex data values. I began to understand that data was so much more than just another component of a complete science project. Even a simple visualization could help illustrate and communicate stories behind research questions, making their implications feel more applicable and relevant to our daily lives. The results of research are certainly not meant to be confined to numbers, simple bars, and straight lines; they are meant to tell a story and express something appealing, compelling, and intuitive in a way that technical phrasing would struggle to achieve. By transforming those uninspiring data tables into something we can understand at a glance, we build for ourselves a tool that is essential to teaching people about ideas. Grueling as it may be, there is certainly an element of art to it.
But, mandatory middle school science fair projects taught us to fill in the wide gaps in our analyses sections with blindingly colorful graphs and images, while our weekly class presentations familiarized us with the monotony of identical bar graphs and the occasional scatter plot reserved for when we feel particularly adventurous. These experiences led us to understand data visualization as a necessary evil, an item in a long checklist of tasks needed to “correctly present science.” Sometimes these requirements yielded mild success, producing poster displays with eye-catching flowers of color. But perhaps more often and unfortunately so, these efforts simply produce cookie-cutter visual reiterations of our boldfaced project titles.
Good data visualization is no easy task to accomplish, and anyone that has worked with an Excel spreadsheet is all too familiar with the array of problems that can arise when building your first visualization. Shift a bar here to give a little more emphasis to the variable in question, and the graph starts to look off-center. Tweak an axis scale in hopes of getting rid of some weird-looking spacing, and all the pretty bars you spent the last hour adjusting to be just the right width vanish before your teary eyes. Every supposed fix brings a new set of problems like some cruel sudoku that was never meant to be solved.
Dr. Monson walked us through solutions to correct several common issues before putting us to the test by asking us to identify examples of bad data visualizations. After completing this exercise, we quickly learned that data visualization is often poorly done not so much because researchers are unable to work with data but rather because they are often unaware of the simple ways to create effective visualizations.
Well-done visualizations bring to attention what began as an investigation. They reveal what components of our culture, society, or even daily life, first prompted a question worth researching. Data visualization is a story telling medium, and a powerful tool that when coupled with a holistic presentation, can be much more than just simple numbers and basic trends.
Next time you need to make a visualization, rather than dread it, take it as a chance to make your own story and to spin a tale. Research’s purpose is to answer our questions about the world in which we live. Data is simply one step in reaching this goal. With the right kind of data visualization, we can better convey our research ideas and more importantly, tell stories about the questions that intrigue us most.
Jason Zhang, Huang Fellow ’18
Jason is interested in studying neuroscience and computer science. He is particularly interested in understanding, and perhaps one day simulating, abstract concepts like memory, thought, personality, and emotion.