top of page

You


We’re all in this dataviz field for different specific reasons, but I’d like to think one commonality among us is the belief in objective truth.

  • Facts are facts.

  • You can’t argue with the numbers.

  • Evidence beats speculation.

  • Cite your sources.

And so on.

And in the course of performing our jobs, we strive to be as faithful to the facts as we possibly can be. We don’t create intentionally misleading interfaces, or conjecture wildly about causality. We let the data speak for itself (or themselves, if you’re a “data is plural” kind of cat).


But the real truth is,

you

are in every assessment you deliver,

every viz you create,

every editorial decision you make.

You come through,

in the choices you know you make,

and in the choices you aren’t even aware of.

 

I wrote once before about “finding your voice.” Inherent in that post was the notion of intentionality—developing your style, your areas of interest, your technical skills and tendencies.

But beyond that intentionality, your voice can, and does, come through your work, in ways you may not even consider.

It’s commonly suggested that we can create products that are meaningful and engaging to the audience by making that design relevant to each viewer. That is to say, let the reader or user find themselves in the data, or make the whole piece pertinent to that specific individual.

But before we even get to that step, we’re unintentionally making the design personal and specific to ourselves.

 

Say you’re provided a data set to explore.

  • Did you filter it?

  • Did you aggregate it?

  • Did you seek out additional datasets to provide context?

  • Where did you look for those?

  • Did you find six additional data sets and discard four of them? Why?

In your first analysis, you found outliers in your data.

  • What did you do with them?

  • Are they illustrative of an insight, or are they the result of random chance and sampling variations?

  • How would you know?

  • Do you plan to highlight the outliers, or exclude them, or just leave them as is?

Now you’re creating a visualization for a client with specific requirements.

  • What relationships are you highlighting?

  • Why?

  • Are you grouping people together?

  • Why are you choosing to group by dimension A instead of dimension B?

  • Are you considering the implications of these groupings?

You picked a palette for your design.

  • Are all the colors of equivalent value, or is one bolder than the rest?

  • Are you assigning red to a certain value?

  • Did you pick the colors to elicit an emotional response, or because they looked pretty?

  • Did you put a break point in a diverging color ramp?

  • Why did you put it where you put it?

Finally you deliver the product to your customer, or to the public. How can the audience interact with it?

  • What insights are easy to derive?

  • What takes several clicks, or requires comparison across multiple screens?

  • What relationships have your interface choices made obvious, and which ones require multiple clicks, or multiple comparisons across screens? Sure, the data is all there, but which connections are people *likely* to make with the product you delivered?

That's a whole lot of you, baked into what is meant to be an objective, data-driven product. Maybe far more than you thought there would be; maybe far more than you had ever realized.

 

Look at any community project for examples of this.

For initiatives like Viz for Social Good, Data For a Cause, Makeover Mondays, or SWD Challenges, dozens of developers start with similar, if not identical, data sets and requirements. But the divergence in finished products is astounding. All of those choices--even the choice not to modify default settings--is an injection of you into the data.

As much as we try to be disinterested (not uninterested) observers when we are creating tools for others to interact with data, each decision we make is a subtle addition of ourselves, and our personal view on the data and the business questions at hand.

 

There are oh, I don't know, a MILLION inspirational-quotes-memes using this old Albert Einstein chestnut:

"The important thing is not to stop questioning."


Believe or not, as a small child, I had this quotation on a tile next to my bed. (In the 70s, you bought your memes from the Hallmark Store and hung them on the wall.)

But it still holds true in this instance. Don't stop questioning, by which I mean: don't stop questioning yourself, about the decisions you're making. If someone asks you, "Why did you decide to do X," you should always have an answer that isn't "Oh, that's the default," or "I don't know, I just like it," or "I never thought about it."

I doubt there are many black-hat Tableau developers out there, using dark dashboarding arts to mislead the public intentionally. But we all make choices that aren't fully considered, whether it's because of time constraints, or a lack of awareness of issues, or assumptions about our data/customer/requirements/whatever.

There’s neither a robotic, automatically-correct answer for every decision, nor a way to abstain from putting ourselves into everything we create. The you--the us--that ends up in our products should be included *specifically* to serve the data and to serve the analysis, in the best ways we see fit.

12 views

Recent Posts

See All
bottom of page