Making Sense of Visualization Literacy

Initial ideas on what visualization literacy is about and what we could do in this space

I have been meaning to write about data literacy for a while, but I have been hesitant because data literacy is one of those overloaded terms that means different things to different people, and I was unsure how to approach it. So, this post is my attempt to reason publicly about what data literacy, specifically data visualization literacy, is. The term is often thrown around, but I have not seen any meaningful and accessible analysis of what it is about and its main objectives.

The broad concept of data literacy has been around for a while, and I sense it’s been taking up speed lately. I see many more professionals working in this space, more events, more companies, more research, etc. Within data literacy, we can include the slightly narrower concept of “visualization literacy,” which focuses specifically on the data visualization part of data literacy. But what is data visualization literacy? What is its scope? What challenges and opportunities exist in this space? In this post, I’ll try to trace its lines and see if we can get a better sense of it.

Defining Visualization Literacy

The interesting fact about data visualization literacy is that giving a definition is not easy. The concepts can be molded in different ways. The only sure thing is that it’s about acquiring abilities in doing things with data visualization. The first big distinction one has to make is the difference between being able to consume versus produce data visualizations. It is certainly true that these two branches have skills in common; however, teaching people to consume visualization appropriately is a more pressing need and one that can and should be scaled as much as possible. In other words, while we do need to empower people with data analysis and presentation skills, there will always be more people who do not need to produce data visualizations directly and yet must be able to consume them effectively. For this reason, in the rest of the post, I will focus only on the consumption side of the equation.

When we focus on abilities in consuming data visualization as a reader, we must categorize these abilities in a hierarchy, which can then be used to organize the skills and design educational interventions. Currently. I see three main levels of abilities:

• Level 1: Basic reading. At this level, a person can identify a chart type (or figure out how to read a chart they do not know yet) and understand how to extract basic information from it. For example, if I show a bar chart, they will be able to recognize it, and they will know that the proper way to read it is to compare the categories represented by the bar according to the values represented by the height of the bars.

• Level 2: Identifying and extracting facts. At this level, a person should be able to make more direct connections between the reality represented by the data and the visualization and extract facts from the representation. For example, if I show a line chart comparing five countries according to some parameter, they will be able to extract facts such as when a country overtakes another, peaks and valleys, inverse trends, and their meaning.

• Level 3: Making inferences and assessing validity. At this level, a person knows how to reason about the reality the data represents and makes inferences that go beyond what is represented. That is, the person will be able to connect the information in the data with their knowledge of the problem and make inferences that go beyond the data. In addition one should be able to assess the validity of a graph and the conclusions extracted out of it to assess their validity.

These levels are a very rough sketch, and they should be defined more precisely, demonstrating that much more rigorous work is needed in this space.

Why Care?

A legitimate question we must answer before moving forward is, “Why do we care about visualization literacy?” We care because studying visualization literacy helps us trace the boundaries of our discipline and systematically structure its content. Once we know its boundaries and its components we can start studying methods to help acquire this knowledge quickly and effectively.

A larger question is why we care about people learning this knowledge. The answer is that in a world that provides increasing opportunities to describe the world through data, knowing how to think with data will be a fundamental skill for everybody. First, developing the critical thinking skills necessary to evaluate the validity of the data presented will be crucial to developing judgment. Imagine being unable to read a plot or understand if a given statistic can’t be true. Second, those who want to use data to answer their questions will be able to assess reality much more deeply. Just imagine being able to geek out with data while looking for the house you want to buy (I have a colleague who went full data scientist on that!).

Once we agree that developing visualization literacy is an important societal goal, the next question is, how do we achieve it?

I like to think about two main components in this space: how to assess and measure the level of literacy and how to improve and develop the level of literacy. Let’s take a look at both.

Measuring Literacy

Measuring visualization literacy is necessary for several reasons. First, if you want to develop and compare pedagogical interventions, being able to measure what and how much a learner has learned is crucial to developing a scientific approach to the problems. Learning scientists need ways to measure to study pedagogical interventions. Second, much learning is happening and will increasingly happen through digital apps and tutors. Measurement and the characterization of the learned knowledge are essential to deciding how to guide a student through a personalized path. Learning apps can speed up or slow down and decide which element to spend more time on to cover existing knowledge gaps. Third, visualization practitioners and researchers need ways to measure literacy, tailor the complexity of their visualizations, and study the effect of different designs while using the literacy level as a controlled variable. There are probably more good reasons to develop good measurement procedures. There are already a few around, and I plan to write a whole new follow-up post focusing only on the existing techniques (spoiler alert: I had the good fortune to be involved in the early development of one - you’ll learn more about it in my future post).

Educational Interventions

This is, for me, the most exciting part. How do we teach people to think with data effectively? There are probably a million different ways, and we need many different ways so that people can use the ones that work best for them. As an educator, I am always torn between three main (non-mutually exclusive) approaches:

• Principles. Teaching by explaining the principles.

• Examples. Teaching by providing examples.

• Experiences. Teaching by giving problems to solve.

The truth is that probably a mix of these approaches is best, even though, in recent years, I have tried to focus more on exposing students to actual problems that are as close to reality as possible.

Another interesting problem is what medium to use to teach people visualization literacy. Some examples are courses (like traditional school and college courses and online courses), books, and applications. As much as I believe we need way more courses and books, I am intrigued by the idea that we could build apps that teach students how to think with data. I spent some time thinking about it and plan to report on it in a future post. Creating apps that walk people through data-rich experiences to teach them how to think effectively with data seems possible.

A good example is “C'est La Vis,” an application developed by researchers a few years ago to study visualization literacy in elementary school. The application and the findings are described in a paper by Alper et al. titled "Visualization literacy at elementary school" and it’s a good example of what is possible to do in this space.¹

Conclusion

Visualization (and data) literacy is one of our field's most important and consequential topics. Understanding what we need to teach and how to improve people’s knowledge and skills in this space is one of the grand challenges of the world of data.

In this post, I started sharing a few ideas about this topic, but I plan to get back to it often. As I align my research towards this topic, I want to ensure I’ll share some of this joinery with you. There is already a lot to share.

In my upcoming posts on this topic, I’d like to cover more existing methods to measure literacy and ideas about building educational apps in this space. Please let me know in the comments below if you have any ideas. I am always happy to learn more about what your thoughts are! Thanks for reading.

1. If you want to listen to a podcast about it, here is an interview with the authors I recorded with Moritz on Data Stories a few years ago.