Data Stories #35: Visual Storytelling w/ Alberto Cairo and Robert Kosara

datastories fun storytelling

Hi all,

Hot topic today! We invited Alberto Cairo and Robert Kosara to discuss the role of storytelling in visualization.  What is storytelling? Is all visualization storytelling? Should we always strive for telling a story? How does storytelling match with exploratory visualization? Should we aim more for worlds and macroscopes than stories as Moritz advocated a while back at Visualized? We went on a somewhat lengthy discussion on these topics and I think we all ended up agreeing on a lot of things and developed a much more nuanced view of storytelling. As you can see from the picture we had lots of fun (thanks Robert for taking the screenshot). Fantastic chat!

Note: Alberto had a lot more to say after the episode so he decided to publish a follow up post that clarifies some of the things he said on the show. But — spoiler alert — listen to the episode first! :)

P.S. Big, big thanks to Fabricio Tavares for taking care of the audio editing of this episode!

Links

14 thoughts on “Data Stories #35: Visual Storytelling w/ Alberto Cairo and Robert Kosara

  1. Moritz, I loved your talk at Visualized and couldn’t agree more with your original position (Worlds, not stories!). I first came across it while working on a data visualization talk for our regional Society of Professional Journalists conference in which I arrived at the same point about worlds (I called them universes).

    I listened to this whole podcast to see if there was anything that I found convincing on the other side and I did not :) But I did have a few other thoughts

    On the topic of the kind of story telling data visualization is and the idea of introducing data visualizations with a part of the story for context: Perhaps the best analogy to the real power of data viz is those choose-your-own-adventure books. The idea that the user can have agency and that the story does not have a pre-determined path, but many simultaneous paths is a fundamentally different kind of storytelling that does not necessarily agree with our hardwired tendencies toward narrative structure. But I think the idea that the possibility of simultaneous narratives for a single topic is an important evolution in how we see the world.

    On the topic of writing for your audience: The general public vs specialists. I think the goal of visualizations is to enable the general public to BECOME specialists, if only for the moment when they’re looking at the visualization.

    In the podcasts, there was a discussion about how people remember the facts in stories better than if they were presented out of a narrative context. This just underscores how faulty memories are. If a complex issue doesn’t fit neatly into a story, are we doomed to just forget it? The power of data visualization is that it gives us a way to compensate for our limited ability to keep facts straight, by simply letting us view all of them in context all at once. But that’s a world, and not a story.

    One final point: as a lover of film and data viz, I think you nailed it when you suggested that visualization may not be best medium for the kind of evocative storytelling often implicated by people who really want to connect emotionally with their audience.

    All of this was a long way of saying, I hope you don’t cede this debate to the pro-storytelling side! The more I hear/think about it, the more convinced I am that “worlds” and not stories are the only legitimate use of data visualization

  2. It’s the end of the semester at the University of Miami, so all my neurons are busy grading and comforting weeping students (just kidding.) I’ll try to be brief.

    First of all, I don’t think that there are “two sides” in this debate. We can pretend that they exist, as discussions of this kind are always fun (Santiago Ortiz and I had one at Tapestry, for instance.)

    I’d like to remind everyone of what’s written here:

    http://www.thefunctionalart.com/2014/04/annotation-narrative-and-storytelling.html

    I don’t see this conversation as a “storytelling yes / storytelling no” fight. I believe that annotating graphics in one way or another is always necessary, yes, but structuring a visualization as a narrative (or several parallel or intertwined narratives) or a story isn’t. It depends on the nature of the data, your audience, your goals, etc.

    And I’m really wary of the word “storytelling,” at least in the sense that I defined in the post above.

    Now, more comments:

    —I hope you don’t cede this debate to the pro-storytelling side! The more I hear/think about it, the more convinced I am that “worlds” and not stories are the only legitimate use of data visualization——

    The “only legitimate use.” Those are strong words. You’re trying to completely rule out several useful tools and techniques. Do it at your own risk. In the meantime, we on “the other side” will keep our minds flexible and open to use whatever device and technique and tool available (narrative, exploration, you name it —even Hans Rosling-style videos) to make the world a more understandable place. That’s truly the “only legitimate use” of visualization.

    ——Perhaps the best analogy to the real power of data viz is those choose-your-own-adventure books. The idea that the user can have agency and that the story does not have a pre-determined path, but many simultaneous paths is a fundamentally different kind of storytelling that does not necessarily agree with our hardwired tendencies toward narrative structure——

    This is a bad analogy. I was a huge fan of those books and, if I remember well, they all began with a compelling short narrative that set the mood and introduced the setting and some characters. Then, they let you choose your own path. So we on “the other side” of this debate could steal this example to make our case.

    ——On the topic of writing for your audience: The general public vs specialists. I think the goal of visualizations is to enable the general public to BECOME specialists, if only for the moment when they’re looking at the visualization.——

    Beautiful words at an abstract level. But take a look at this beautiful visualization of languages used on Twitter all over Europe:

    http://www.fastcodesign.com/1665366/infographic-of-the-day-the-many-languages-of-twitter

    Assume that we make it interactive, letting people zooming in and out, etc. In what sense does it make me a “specialist”? I don’t think it does, or it doesn’t do it enough. It leaves too many interesting bits of information go unexplained, too many questions are unanswered.

    Is it a bad visualization? Not at all. It’s amazing. But if you design it for me (a general viewer/reader,) I need some guidance to extract meaning from it. I need annotation, at least. I can see many people in those countries who are not tweeting in the national languages. Who are those people? Why are they there? I want to know.

    These maps are already fantastic. With some extra context —provided by annotation, highlighting, or even narrative— they could be superb.

    —If a complex issue doesn’t fit neatly into a story, are we doomed to just forget it?—

    No. If a complex issue doesn’t fit into a story, you simply don’t use a story as a means to convey your data. You do something else. I think that this is something that we made clear.

    ——The power of data visualization is that it gives us a way to compensate for our limited ability to keep facts straight, by simply letting us view all of them in context all at once. But that’s a world, and not a story.——

    Why not offering me both, when it’s possible and appropriate?

    • Great comments all around! I appreciate the opportunity to engage in such a discussion. Let me first say that I recognize that everyone featured in the podcast has almost certainly spent more time thinking through and about all manner of issues related to data visualization, and it shows in the depth, sophistication, and expert articulation of the commentary.

      Nevertheless, I feel strongly about the topic and thought I should offer my opinions, half formed as they may be :)

      – I don’t see this conversation as a “storytelling yes / storytelling no” fight.
      I would agree with this. I read Moritz’s original post not as a commentary on the inherent usefuless of storytelling with data, but rather proposing that data “worlds” are inherently more valuable than data “stories” (however you choose to define them). Moreover “worlds” exploit the unique advantages of data visualization as a medium of communication, that stories cannot (in my mind information density). The resulting conclusion, “Worlds, not stories” then follows the logic if Worlds > Stories, why spend your time constructing stories?

      I do believe that some stories (or perhaps narratives in the definition from your letter which I think is quite good) can best be told with data, however, I think they are much less valuable without the “world” of supporting data to provide a more complete context. So even stories (which as you point out are essentially optional structures on top of data visualization) benefit from the presence and prioritization of worlds. Ultimately, I would suggest if a narrative / story is required or desired for the audience, that it be layered on top of a “world” that is also presented

      – The “only legitimate use.” Those are strong words. You’re trying to completely rule out several useful tools and techniques. Do it at your own risk. In the meantime, we on “the other side” will keep our minds flexible and open to use whatever device and technique and tool available (narrative, exploration, you name it —even Hans Rosling-style videos) to make the world a more understandable place. That’s truly the “only legitimate use” of visualization.

      Haha, I apologize, I’m prone to hyperbole. I like to take strong positions because at least then there’s something to talk about :) But in the end, perhaps you’re falling victim to the same lack of specificity in your choice of words. Is “mak[ing] the world a more understandable place” a completely legitimate and appropriate use of data visualization? Stories about magic, mythology, religion, and stereotypes make the world a more understandable place. Is that really how we want people to think about data visualization? I would think that statement needs something about “approaching truth” as well as facilitating understanding

      – This is a bad analogy. I was a huge fan of those books and, if I remember well, they all began with a compelling short narrative that set the mood and introduced the setting and some characters. Then, they let you choose your own path. So we on “the other side” of this debate could steal this example to make our case.

      Oh, I was actually trying to make a sort of concession, precisely for the reason you stated. Setting the mood and creating a compelling narrative (story techniques) to set the stage for exploration, may be an essential part of effective data visualization, too. Once you start the choose your own adventure portion, though, you start to depart from traditional storytelling and narrative techniques, and I think that’s important.

      – Assume that we make it interactive, letting people zooming in and out, etc. In what sense does it make me a “specialist”?

      What is a specialist, other than someone with domain-level knowledge and expertise that understands the big picture and knows where to look next for the next piece of information? I think data visualizations have the capacity to bring people to this level while they are interacting with the viz.

      – No. If a complex issue doesn’t fit into a story, you simply don’t use a story as a means to convey your data. You do something else. I think that this is something that we made clear.

      We may simply disagree on this matter. What I mean by that, is I don’t see stories as one of several equivalent options for communicating data. For better or worse, I think of stories as this necessary evil, this thing you do when you don’t have a better or more complete option. In the world of communicating things, I think stories, like statistical models, are always incomplete but the the simplification serves a specific purpose: making things memorable and understandable (accessing the way we process sequences, cause and effect). Exploration-oriented “worlds” Data visualization strikes me as having a different strength, something closer to “intuition” or the stuff Malcolm Gladwell wrote about in his book Blink. Through exploring the data, we can subconsciously develop an understanding of it through observation, before we have to construct a narrative about how it works. And I’d argue that intuitive understanding has the potential to be more useful than the narrative, which takes that intuition and flattens it into a simplified sequence that might miss important dynamics.

      ——The power of data visualization is that it gives us a way to compensate for our limited ability to keep facts straight, by simply letting us view all of them in context all at once. But that’s a world, and not a story.——

      – Why not offering me both, when it’s possible and appropriate?
      Let me offer this as an evolution of my original position. I understand the value of stories in certain cases, but agree that they are not always necessary. Perhaps, what I’m saying is that “worlds”, like annotation are a necessary part of (good) data visualization, where stories are not. You need worlds when you tell stories with data, because they provide the context you need to know whether you should believe the data or not.

      (and since we made the earlier comparison to film, this passes that check as well. The best films are the ones that have fleshed out the worlds the characters live in. That’s why we believe those worlds, rate some as better than others and why the stories within them have more impact)

      Thanks again for this wonderful discussion. I’ve been thinking about it a lot recently and I appreciate the opportunity to participate and have my ideas challenged and evaluated by the best in the field!

      • Well, thank you. Just two minor things:

        —-I would suggest if a narrative / story is required or desired for the audience, that it be layered on top of a “world” that is also presented—-

        This has been my point all along. Why not use both a presentation layer (which could be based on annotation or, in some cases, a narrative) and an exploration layer?

        —-Is “mak[ing] the world a more understandable place” a completely legitimate and appropriate use of data visualization?—–

        Yes, it is. I believe that it should be our main guiding ethical principle, actually.

        ——Stories about magic, mythology, religion, and stereotypes make the world a more understandable place.—

        No, they don’t. Quite the contrary. They usually muddle understanding. I was obviously referring to evidence-based strategies. “Approaching truth” was implied.

  3. Really great podcast! I particularly liked the discussion around distinction between annotation, narration, and story and the “so what” test. Thanks for taking the time to discuss and publish this!

    • The study is tricky though. It is very well executed but it tells nothing about whether people remember the content or message of the chart and how accurately they remember it. It’s more a study on recognition than else.

  4. Great podcast. I wrote up a short reflection on my blog at http://neoformix.com/2014/DataStories.html .

    key excerpt:
    …I’m suspicious of salesmanship and marketing. I’m wary of other people using emotion and a good story to persuade me to believe something that isn’t true. I have some concern that data visualization work that emphasizes storytelling is more likely to be ‘Data Fiction’ – or propaganda. The designer, through careful choice of selected facts, use of emotion, drama, conflict, and all the other techniques of storytelling can craft a message at odds with reality. The use of ‘data’ will even lend an air of authority to that message.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>