LOPEZ Group employees dropped everything and learned in First Philippine Holdings Corporation’s (FPH) Friday DEAL (Drop Everything and Learn) for March 31, “The Art of Data Storytelling.”
Launched in February 2022, Friday DEAL seeks to spur new ideas, learn new ways of seeing and doing things, and unfold new skills through bite-sized learning sessions. “The Art of Data Storytelling” hosted by FPH-First Gen HR training officer Jun Prila was the third one for 2023 and serves as the follow-up to “Data Science 101” in January.
Gerald Valentin, lead trainer at consulting company StoryIQ, returned for the sequel.
The nearly 300 attendees included employees and guests from FPH, First Gen, Pi Health, Energy Development Corporation, First Philippine Industrial Park, Infopro Business Solutions Inc., Pi Energy, Knowledge Channel Foundation Inc., First Gen Energy Solutions Inc., ABS-CBN, Lopez Holdings Corporation, First Philec, First Philippine Industrial Corporation, INAEC, Rockwell, ThermaPrime and Third Generation Holdings Corporation.
Visuals plus narrative
So what is data storytelling? The participants related it to “persuasion,” “analysis,” “insights,” “visualization” and “interpretation.”
According to Valentin, data storytelling is “basically the same as data communication,” where one makes use of data visualization principles and storytelling techniques to put their message across to the audience. The ideal amalgamation of visuals and narrative can lessen the cognitive load on the audience and allow them to make the best decisions.
Cognitive load refers to how the brain processes things around it. Ever wonder why you feel tired after a day of working on your computer? That’s because your brain is constantly, endlessly busy—you might be answering emails or chats, downloading files, looking something up on Google and drafting reports at the same time—while carving out time for a Zoom meeting or engaging with colleagues if in the office.
Using up all of your energy and mind power results in ego depletion, making you feel tired or drained even if you are just stationed in front of your computer. When people are cognitively depleted, bad things happen; that is, they are more likely to make bad decisions, Valentin said.
The goal, then, is to lessen the cognitive load of the audience so that at the end of the meeting, they would still have the energy to think and make the proper decision.
“If they were too exhausted in trying to understand your visuals, they would not have enough energy to make the right decision. The question is, how?,” Valentin said.
He started out with an initial build that he polished and improved on throughout the session, guided by the insights gleaned from the tests and quizzes answered by the participants and the 4Ds of data storytelling: define, display, declutter and direct.
Define your key takeaway so that it is obvious to your audience why they are looking at the visual and why they should care about this visual.
Font size hierarchy
Valentin recommended several techniques, including placing the key message on the “prime real estate” on the slide (that is, the upper left-hand corner) and using font size hierarchy (pick the biggest font size for the main headline). To further refine the message, one may make use of alliteration, repetition, rhymes and affective language or impactful words.
The next step is to improve the display itself, Valentin said, stressing that the display and main message should convey a single message. He cited a 1984 study on graphical perception by statisticians William S. Cleveland and Robert McGill that ranked elementary perceptual tasks, which indicated it is hardest to immediately see the values of visuals based on hue or color and volume, making them least accurate; those based on position (nonaligned scale and common scale) are the most accurate.
“The study shows charts are not created equally. Some are harder to read than others. Our goal when we choose a visualization is to choose one that’s easy to understand,” Valentin said.
This is not to say, however, that one should always choose positions on a common scale. A pie chart may be used to show proportion, composition or parts of a whole, for instance, although care must be taken as it can be tricky to perceive data encoded on such, Valentin advised.
Decluttering the chart and removing extraneous elements will further spotlight the data. “Above all, show the data,” Valentin said, quoting statistician and data visualization pioneer Edward R. Tufte. “Chart junk” may include grid lines, unneeded y-axis elements and legends (“provide cognitive load because the audience has to remember the information” but may be replaced with data labels).
Odd one out
Lastly, direct the eyes or grab the attention of the audience by changing or applying preattentive attributes, such as color, size, line width, orientation, length, shape and curvature of various features, or adding marks. This was implemented on the sample chart by muting the colors of the other categories, and using the main category’s brand color while increasing its font size.
“It’s always the odd one out. We always see the different ones,” Valentin explained.
Comparing the initial build with the new version, he underscored that the latter is easier to understand: “Now we have vertical flow: the title or headline is working hand in hand with the visual. The key takeaway is obvious, thus lessening the cognitive load on your audience. That is our goal whenever we do a data storytelling presentation.”
In their comments and feedback, participants shared their thoughts about Friday DEAL: “I will definitely use the 4Ds in presenting data/information,” wrote an instrumentation and control engineer.
An HR specialist observed: “Lessening cognitive load is the main objective/target in creating effective data presentations… The visual examples were also an effective manner for us to learn—not only did the instructor tell us what effective data storytelling is, he actually showed us how it really looks.”
Reiterating Valentin’s message on cognitive load, First Gen data scientist Clowee Anne Licsi noted that technical people tend to include a lot of information on their decks because they understand it. But the audience might not, so it is important to simplify the data and be intentional in using chart elements, such as colors and grid lines.
“As much as possible remove the noise and you’ll be surprised… your slides or your charts will actually look cleaner,” she said.
Licsi, who used to be in HR, shared her testimony as a participant in an earlier course on data storytelling. She recalled that the team had leveraged the lessons and concepts they learned in the course for their work in HR.
“What I learned in 2019, I’m still using it now when I transitioned to the data science team. You’ll be surprised by how you can effectively communicate visually,” she said.