Katie Fitzenreiter

One of the most effective ways that researchers can communicate their findings is through data visualization. Emily Nastase said it best; she described data visualization as a way to effectively reveal trends, tell a story, and represent things that are not easily observed from words and numbers alone. However, producing a visual representation of data can be quite challenging.

As scientists, we collect complex pools of data and spend countless hours analyzing it. We become so engrossed in our own research that we sometimes forget how obscure our subject matter is to most people who are unfamiliar with our work. Practicing data visualization techniques that we have learned in class can help us forge stronger lines of communication with people outside of our fields, and help us to better understand our own data as well. According to David McCandless, a world-renowned data journalist and information designer, in his Ted Talk, "data is the new soil." This means that data is a "fertile creative medium" from which "visualizations and infographics are blooming."

David McCandless presents examples of figures and graphs that creatively and humorously tell a story based on statistical data. Posted by TED-Ed under Standard YouTube License

The way we present our data in a figure, graph, or other visual display depends on our audience; therefore, knowing your audience is the most important thing to consider when creating visualizations. The second most important thing is knowing your data. It is helpful to ask yourself, "What is my data telling me, and what is the significance?" Knowing your data is crucial to communicating it clearly and efficiently. In order to tell a captivating story of your data, you need to know the information inside and out. After all, how can your audience understand your data if you don't understand it yourself?

Tips for creating your visualization:

• Know your audience
• Know your data
• Find that story to tell (captivate your audience, don't lose them)
• Let the data do the talking (data should stand alone without a caption)
• Eliminate variation in font style, font color, and font size (consistency is key)
• Avoid "chart junk" (unnecessary gridlines, graphics, or anything distracting from the data itself)
• Red and green are "loaded" colors (use them sparingly if you must use them at all)
• Include an active title - clearly and effectively states what your data shows
• Sites and stations should be given descriptive and meaningful labels (i.e. based on location, depth, zone type, etc.). Avoid "meaningless" letters and numbers (i.e. Site A, Site B, Site 1, Site 2)
• Go with the largest font size possible (without drawing attention from the data)
• Stay away from "Acronym Hell" - too many acronyms confuse the viewer
• Avoid using colors that are too closely related (red and pink lines on a graph are a "no-no")
• For colorblind viewers, try to use different shapes if they struggle to distinguish between colors of symbols

I learned a lot in class last week about what it takes to clearly and effectively present your data in a visual display. Below is a figure I had created before the class. Can you guess what I did wrong with it?

Where I went wrong:

• I should have used a stick plot instead of a North and East velocity plot (would have been easier to understand and interpret)
• Too much white space in my plot made my figure look static and uninteresting
• The legend should have been larger
• I could have increased the font size
• I failed to include wind velocity data to compare it to drifter velocity

On the other hand, this figure, produced by the Chesapeake Bay Stock Assessment Committee is an excellent example of how statistical data is used to tell a story and reveal trends clearly and efficiently:

Here's what the author(s) did correctly:

• Allowed the data to stand alone and speak for itself
• Kept font size, font style, and font color consistent
• Used colors that were easily distinguishable from one another
• Included a legend that was easy to read
• Data tells a story and reveals trends
• Avoided chart junk

There is no title for the figures, but I think the data speaks for itself well enough that a title is not needed in this case.

Overall, applying data visualization techniques is useful for people in all stages of their scientific career. It is never too late nor too early to improve your methods of data communication to a selected audience, and there is always room for improvement. When presenting data, always remember to modify your presentation based on your audience, and be sure to ask yourself, "What is my data telling me?" Lastly, when communicating research through a visual presentation, your data should always be the centerpiece.

¹ "North America Total Primary Energy Consumption in 2015" by Shahzad under CC BY-SA 4.0

From the Chesapeake Bay Stock Assessment Committee, 2012 Blue Crab Advisory Report Figures [pdf].