Introduction to Rhetorical Data Visualization

Introducing my new course on the rhetorical aspects of Data Visualization.

I just started a new course! It’s called Rhetorical Data Visualization. You can watch the first introductory lecture on Loom by clicking the link below.

RhetVis L01: Introduction to Rhetorical Data Visualization - Watch Video

Click here to watch the 1st video lecture on Rhetorical Data Visualization.

I will use this newsletter to report about the course, to share the material I produce, and the thinking generated by developing and delivering the course.

What is Rhetorical Data Visualization, and why did I decide to design this course? The course stems from the realization that every data visualization (and every type of communication with data, being of a graphical nature or not) has a specific framing that makes certain interpretations more likely than others. Hullman and Diakopoulos have brilliantly proposed this idea in their landmark paper in Visualization Rhetoric.

In a way, all visualizations are somewhat biased. They reflect a specific angle. One may be tempted to believe that “data is truth” and that if you are “neutral,” then things will be fine. I don’t think this neutrality exists, and I am not even sure it’s desirable (more about this in a moment). The truth is that every time we use data in support of an argument, we must make choices that reflect our inclinations. At the very minimum, we have questions we focus on, variables and data items we decide to include or exclude, and graphical representations we decide to use. It’s not like we can NOT do these things. You have to make choices, and choices represent your inclinations.

At the same time, we can’t throw our hands in the air and give up. Does this mean that all this data thing is futile? Not at all. Evidence is useful. Good evidence is super useful. And good arguments are super, super useful! We just need to realize everything comes from a perspective.

However, how do I teach that? What do I include in a course on the rhetoric of visualization? I’ve been thinking about this the whole summer, and I came up with a first approximation stemming from the question, “What skills do I want students to have after taking this course?” The answer is that I want the students to have evaluative and generative power; that is, they should be able to analyze existing visualizations under the lens of rhetoric, and they should be able to build rhetorical visualization mindfully from a place of humility, curiosity and, above all, integrity.

The course is (tentatively) organized around these modules.

1. Data-reality gaps. We have mental models of what data represents, and often, there is a mismatch between what we think it represents and what it represents.

2. Data transformations. We almost never consume data “raw.” We select variables, aggregate and filter items, compute statistics, etc. All of these bring a “perspective.”

3. Statistics and statistical fallacies. Data often become statistics. But statistics can often lead to erroneous interpretations and inferences and just be plain wrong. Being able to recognize problems with statistics is a master skill. One that admittedly can make you very humble because you feel like there is always a new way to get fooled. But covering the basics is already a big step.

4. Graphical representations. For any given piece of information, there are innumerable ways to represent it. Some are more intuitive than others, some can skew attention or interpretation more to one fact than another, and others are suspiciously misleading. Understanding how this works is crucial for reading, evaluating, and producing visualizations.

5. Contextual elements, narrative, and delivery mechanisms. We are used to talking about how to map data to shapes, colors, specific types of symbols, layout strategies, etc.. Still, the reality is that the contextual elements of a visualization have a significant effect on interpretation. Titles, annotations, captions, and narrative sequence play a significant role. Then, visualizations can be delivered through many different media and means, such as newspaper articles, scientific papers, presentations with slides and animations, movies, and videos. All of these channels can bring additional layers that greatly impact interpretation.

6. Visualization psychology. It’s helpful to understand the cognitive aspects of data visualization better to understand how people interpret data visualizations. How do people make sense of visualization? How does visualization impact the way people think about something?

7. Ethics and integrity. Since visualization can influence people’s beliefs, it’s crucial to understand its ethical aspects and implications. Similarly, developing a sense of integrity is necessary to build visualizations that try to get as close to the truth as possible.

8. Data dialogues. How do we enable difficult conversations based on data? People often have different perspectives, and data may not have the ultimate answer. How do we learn to disagree respectfully and hopefully learn from each other even when we disagree?

I will publish updates on the course, including links to recorded lectures here. If you don’t want to miss these updates and are not signed to this newsletter yet, sign up!

—

Can I ask you a favor? Please share a link to this post (or the newsletter) on your favorite social media platform if you like what I write. You can also send it directly to a friend or colleague if you know anyone interested. That would help immensely.

Thanks!