An excellent visualization, according to Edward Tufte, expresses "complex ideascommunicated with clarity, precision and efficiency." I would add that anexcellent visualization also tells a story through the graphical depiction ofstatistical information. As I discussed in an earlier post, visualization in its educational orconfirmational role is really a dynamic form of persuasion. Few forms ofcommunication are as persuasive as a compelling narrative. To this end, thevisualization needs to tell a story to the audience. Storytelling helps theviewer gain insight from the data. (For a great example, how much do you think steroids have influencedbaseball?)

So how does a visual designer tell a story with a visualization? The analysishas to find the story that the data supports. Traditional journalism does thisall the time, and journalists have become very good at storytelling withvisualization via infographics. In that vein, here are some journalisticstrategies on telling a good story that apply to data visualizations as well.

• Find the     compelling narrative. Along with     giving an accountof the facts and establishing the connections between them, don't be     boring. You are competing for the viewer's time and attention, so make     sure the narrative has a hook, momentum, or a captivating purpose. Finding     the narrative structure will help you decide whether you actually have a     story to tell. If you don't, then perhaps this visualization should     support exploratory data analysis (EDA) rather than convey information.     However, for the designer of an exploratory visualization it is still     important to spark the viewers' imagination to encourage examining relationships among and facilitate interacting with the data - think     gameification.
• Think about     your audience. What does     the audience know about the topic? Is it meant for decision makers,     general interested parties, or others? The visualization needs to be     framed around the level of information the audience already has, correct     and incorrect:
  
  • Novice: first exposure to the subject, but doesn't      want oversimplification
  • Generalist:      aware of the topic, but looking for an overview understanding and major      themes
  • Managerial: in-depth, actionable understanding of      intricacies and interrelationships with access to detail
  • Expert: more exploration and discovery and less      storytelling with great detail
  • Executive: only has time to glean the significance and      conclusions of weighted probabilities
    
• Be objective     and offer balance. A     visualization should be devoid of bias. Even if it is arguing to influence, it     should be based upon what the data says--not what you want it to say.     Tufte found numerous charts that misled viewers about the underlying data,     and created a formula to quantify such a misleading graphic called the     "Lie Factor." The Lie Factor is equivalent to the     size of the effect shown in the graphic, divided by the size of the effect     in the data. Sometimes it is unintentional-a number that is three times     bigger than another will be perceived nine times bigger if represented in 3D. There are simple ways to encourage objectivity:     labeling to avoid ambiguity, have graphic dimensions match data     dimensions, using standardized units, and keeping design elements from compromising the data.     Balance can come from alternative representations (multiple clustering's;     confidence intervals instead of lines; changing timelines; alternative     color palettes and assignments; variable scaling) of the data in the same     visualization. Maintaining objectivity and balance is not a trivial effort and is easily unintentionally violated.     Viewers and decision makers will eventually sniff out inconsistencies     which in turn will cause the designer to lose trust and credibility, no     matter how good the story.
• Don't Censor. Don't be selective about the data you include or     exclude, unless you're confident you're giving your audience the best     representation of what the data "says". This selectivity     includes using discrete     values when the data is continuous; how you deal with missing,     outlier and out of range values; arbitrary temporal ranges; capped values,     volumes, ranges, and intervals. Viewers will eventually figure that out     and lose trust in the visualization (and any others you might produce).
• Finally,     Edit, Edit, Edit. Also, take     care to really try to explain the data, not just decorate it. Don't fall     into "it looks cool" trap, when it might not be the best way     explain the data. As journalists and writers know, if you are spending more     time editing and improving your visualization than creating it, you are     probably doing something right.