« « Data Delver: Mo Tamman, Wall Street Journal

Visualizing networking: When it doesn’t work » »

Visual confections are more than mere presentation

Data visualization.  It’s one of those terms that can mean so many things.  I say I’m doing an independent study on data visualization this quarter.  That’s true.  But a better description would be “data visualizations for journalism.”  I’ve talked about this with people before, a lot of pieces are gorgeous, and they do convey information, but they don’t tell a story that informs the viewer in a useful way.  Sure, the colors used in Flickr photos is data, and if you represent it visually, that is a visualization.  But that doesn’t make it appropriate for a news Web site.   It’s the same reason simple transcription doesn’t make up all of our news — and if  it ever does, then we really need to save journalism.  Whether through text or words, we use journalism to tell a story.

It’s something I’ve been wrestling with all quarter. It’s the reason I have such a problem with uncontextualized data centers as the primary data initiative in news organizations.  Which is why I found myself saying, “Yes, yes, yes!” to “Visual Confections,” the final chapter I read this week in Edward Tufte’s Visual Explanations.

I love the word confection, and the images it draws to mind.  One definition is of a food rich in sugar.  It’s sweet, pleasurable to the senses.  I always think of confections as more carefully crafted sweets, something I would get in the family-owned fudge shop by my middle school as opposed to a Hershey bar I might see in the checkout line.  And our journalistic creations should be confections in this sense, they must delight as well as inform.  But a second definition of confection, according to Princeton’s WordNet, is concotion, “the act of creating something by compounding or mixing a variety of components.”

And that’s where it gets interesting.   Journalists have the capability to aggregate, analyze and redisplay information.  Even if it already exists on the Web, and the user can access it, we provide the context.  This speaks to me as various professors and editors have referred to some of my more adventurous news projects as “Web concoctions.”

It’s the only chapter in the book where I can recall an example from journalism being used.  The idea of visual confections is that while many visualizations are just representing a data set, you can use a visualization to tell a story, combining different sources and using bits of text — or “chatter” to inform the viewer.

Graphs may approach this when placed next to a story, or if ample text is included with the graph, or if further explanation appears on rollover or click.  And perhaps in exceptional circumstances, not one word of text will be needed, but I think that’s extremely difficult.  It should be reasonably quick for viewers to understand the point of the image or interactive experience, and in most cases, some words will be needed to help facilitate that.

I admired the example the book pulled that accompanied a June 1985 article in the Washington Post, warning viewers about the dangers of the fast-moving Potomac river.  What could have been presented as a boring list of safety tips becomes almost cartoonish in its illustration.  It doesn’t seem silly, but rather, makes the illustration more personable.  The various illustrations within the larger piece range from showing the structure of the waterways, to how you should arrange your body to try to survive the current.  Little blocks of illustration and chatter use Tufte’s principle of several small multiples being more clear than an illustration trying to do everything at once.

Treating a visualization as a confection and creative opportunity does more for grabbing a viewer’s attention, as well.  And it demands more creativity from the creator, which makes it more fun to design, I think.  That enjoyment is passed on to the user, the information sticks in the user’s head.  With that principle in mind, I can now see that a Django-based application, or any data-driven app, is a data visualization by some other name.

Visualizations are much more than how to make that Excel graph easier to understand and less boring — they tell a story that grabs the reader and helps him or her comprehend a diverse set of information.  It’s an attitude that we can bring to journalism of all sorts to take the profession away from stenography and into analysis.

Journalists are a lot like college professors, we must spend the time doing research to discover new things.  But we must also make sure we can convey that diverse set of information to people of all skills and interest levels, just as professors must communicate with everyone from freshmen in their intro class with vague interests to advanced graduate students working on a dissertation who demand very specific information.

« « Data Delver: Mo Tamman, Wall Street Journal

Visualizing networking: When it doesn’t work » »
  • http://www.tamingoftheskew.com Jeremy

    To distinguish between visual confections and journalistic presentations of data, I fall back on the same standard that I use for distinguishing between mere information and valuable journalism: Valuable journalism informs the decisions of its audience members. It’s not just about a “storytelling” (an unfortunate word choice I’ve come to accept I’ll never get journalists to stop using in reference to the task of putting available information in any format that is both relevant and digestible for a particular target audience). In military intelligence terms, it’s about showing how the data are actionable (and therefore worth publishing, i.e. “valuable”).

    As I tell student journalists I speak with, it isn’t news if it doesn’t answer all three of “What?”, “So what?”, and “Now what?” Most charts in the news do OK answering the first one; most charts in the news also fail miserably answering the second and third.

    [Reply]