Striving for a Productive Data Viz Critique

The Data Visualization Society members offer their advice…

Original illustrations in this article are by Martin Telefont

Critique can be a brutal experience. Receiving it and giving it is full of risk and nuance, especially for the untrained.

Because of the diverse backgrounds found among people creating data visualization, the chances are high that you’ll work with someone who wasn’t trained to give or receive critique.

I’m in the never-been-trained camp, so my ears perked up when the bi-weekly discussion topic in the Data Visualization Society Slack channel was “critique —how do we best give it and receive it?”

When I first started creating visualizations, I’d panic at the thought of getting critiqued. It felt like anything was fair game, so I didn’t know what to expect.

But I’ve since learned two very important things:

1. Not all critique is created equal (i.e. it’s reasonable to consider the perspective and knowledge of the giver before absorbing the critique).

2. Growth is uncomfortable, and it always includes some form of critique (so bring it on!).

Knowing this helped me embrace critique, and the community discussion gave me a fuller picture of what a productive critique looks like. Below, I had the honor to summarize the thoughts of these generous, experienced practitioners.

Please leave a comment if we missed any of those aforementioned nuances. ;) And join us in the Slack channel!


Defining Critique

Critique is an analytical evaluation of what works and what doesn’t.

Open question: how is it different than “feedback”? Is that just general impressions or opinions of your work, while “critique” is a formal review of design choices?

Preparation

As the receiver…

First, get in the right mindset. Critique is valuable and helps you grow, so decide to seek it out. We all have a good amount of impostor syndrome, which makes critique feel personal. But if we want data viz to move forward as a field and grow personally, then we need to prioritize receiving critique. Finding people you trust or a mentor can help with impostor syndrome/sensitivity.

Next, gather your “cast of characters”, which can include

  • you as the creator,
  • future you (who has fresh eyes),
  • other data viz practitioners (they understand the struggles),
  • trusted mentors/colleagues (they understand your constraints),
  • someone unfamiliar with your work (ideally from your target audience).

It’s important to recognize which character you’re talking to during the critique (e.g. does this person really know which chart type would work better?).

As the giver:

First, consider if you should ask permission to critique the work.

Open question: Do you always need to ask permission to give critique? On one hand, it shows respect to the creator and increases the likelihood they’ll listen openly. On the other hand, it can limit productive conversations and it might harm the public by letting it go.

During the Critique

As the receiver:

As you engage your cast of characters, state your objective for the critique. Make sure you keep the critique relevant and useful by first defining what you’re looking for critique on (e.g. Who is the audience? What do you want them to know? Is there a specific element that you want feedback on? Or don’t want feedback on?).

Fight the temptation to be defensive. Rather than explain yourself, try to lead with “tell me more.” Even rude or terse critique can be useful, and it’d be a shame to lose the chance to improve because you’re busy responding or taking it personally.

As the giver:

Before giving your assessment of the work, state the intention of your critique. What are you hoping to improve and why?

Open question: Is this necessary if the receiver has already set the objectives?

Many people emphasized that how you communicate is often more important than what you’re communicating. Here are some hard-won tips for delivering your critique:

  • Try “I like…I wish…I wonder…” which is an approach developed in Stanford’s d.school to explaining in detail the intuitive response you have to a work. It requires that you start all feedback with either “I like…”, “I wish…” or “I wonder…” to produce a structured dynamic between giver and receiver.
  • State your suggestion as a question for the receiver to take or leave.
  • Try to give critique in person (or with audio) so the receiver can hear your tone.

In the collective experience, critiques are much more productive when the giver is as specific as possible. Rather than saying, “this could be confusing,” try stating what is specifically confusing to you. A useful exercise is to talk through how you’re experiencing the visualization and weigh the pros and cons of each design compromise. 

Should you do a redesign as a critique?

  • Pro: Easier to show what you mean and conveys that you’re invested in the solution.
  • Con: It can be very time-consuming.

This approach is described in detail in this article by Fernanda Viégas and Martin Wattenberg. If you decide to do a redesign, make sure you state your goal of the redesign and any data simplifications that you performed. Be kind and open with the creator (especially with newbies).

Interesting ideas:

  • Try a “backward” redesign where you redesign a viz and remove key elements that were effective to show how effective they were (another great idea from Viégas and Wattenberg).
  • Do a critique of your own, old work! Remember that one of those people at the table above was your future self, and developing a habit of critiquing your former work will not only make you better at critique, it will make you more comfortable receiving it.

This conversation about productive critique has helped me see the dynamic between giver and receiver more clearly and also notice the nuances of the relationship. I hope this has given you the courage and some tools to execute more productive critiques in your work.


Join the Data Visualization Society for more data viz conversations: https://www.datavisualizationsociety.com/join

Many, many thanks to Martin Telefont for the wonderful illustrations, Mara Averick and Elijah Meeks for your keen editing, and the members of the Data Visualization Society for sharing your expertise.