Learning data visualization

Data visualization sits at the intersection of science and art. My father was a painter and art teacher, which had a big influence on me since the childhood. Years later, this turned into a passion for learning data visualization.

If you work with data, at some point you will need to visualize it. And like most of us, you will open your favorite tool and start creating tables and charts, maybe not making conscious decisions about things like geometry and other pre-attentive attributes. While in some cases this might be fine, it is worth taking a closer look and re-examining your knowledge of data visualization practices, and what you can easily do to improve your skills. In this article I discuss the path I have followed, and I hope it will help you as well.

What you might not know — but need to learn

As data scientists, we are often presented with a multitude of tasks: Finding reliable sources, writing and re-writing queries, doing exploratory data analysis, detecting anomalies, performing statistical testing and predictions — the list goes on. But the outcome of this work is powerful: Presenting insights from data to enable action.

The people who need to take actions are often business decision-makers. The decisions they make because of your insights can make your company prosper, your customers succeed, and your bosses happy. It also gives you the satisfaction of knowing your effort was worth it.

Some data scientists might consider data visualization a nice-to-have. They might believe they have no time to refine annotations in a scatterplot or think about where to place the legend. But if you don’t consider how to present your work, everything you’ve put into it might be diminished by not being heard as you intended. This is your opportunity to be heard, and for your work to make an impact.

My journey into data visualization

Because of my family history and the influence of art, I always had an affinity for drawing and an interest in design. As a data scientist, using proper shapes and colors with structured layouts to communicate important information was something I realized that I wanted to do in my career.

In 2015, I completed a certificate program at the University of Washington called “Business Intelligence: Techniques for Decision Making.” Data visualization was the subject of one of the classes, a relatively new area of focus at the time in academic circles. Although introductory, it gave me insight into a subject that would soon become increasingly popular, and data visualization is one of the skills I now recognize as essential.

Since taking the class, I have learned to pause and ask myself: Am I doing the best I can with communicating data? Is there something more I need to learn and implement before I call my work done?

I have built on that early coursework by immersing myself in literature, blogs, online tutorials, events, and even more formal education. As I’ve progressed, the gaps have become even more obvious. They have helped me understand where I could do better and opened the door for more growth. What follows are my suggestions for your own data visualization journey.

Define your style: Learn and apply knowledge in the real world

Here are three ways I recommend for you to improve your communications, down to the last mile of presenting your work.

Develop an eye for design

For standard reports and interactive tools, what can you do to make them easier to use? Can your audience find what they are looking for?

Reports, dashboards, and scorecards are the core pillars of traditional business intelligence and rhythm-of-business analysis. They contain important KPIs and charts, sometimes separated among different tabs or views. User interaction is usually required to select, sort, and filter. This enables exploratory analysis: Users might not know exactly what they are looking for, and so they spend time looking around, uncovering insights.

For research papers and pre-packaged presentations, ask whether they tell a story and provide any ‘aha’ moments. Do they resonate with your audience, and suggest the next best action?

This is the realm of explanatory analysis and artifacts, which solve specific business problems or answer specific questions. No interactivity is available to users, just a collection of visuals and narrated text. Analysts may need to present these to the stakeholders or have someone else to do that for them. In each case, they need to ensure that the story and key takeaways are clearly understood.

Consider the following chart, which reflects typical default settings in Power BI. We’ll return to it later in this article, after considering some ways to learn how to improve it.

Learn from experts

Capitalize on early visualization thinkers. Edward Tufte and Stephen Few are prominent pioneers in data visualization. Perhaps you’ve heard of them, or even read their books or attended a seminar. Their guidelines have laid a foundation for data visualization practitioners across many fields and industries.

Consider what you might have learned from them already that you could apply to your recent presentations. Was there a moment you could have used a checklist with your visualizations to help them meet some basic criteria of effective design? Next time you’re asked “What does this chart mean?” or “What does the y-axis represent?” it might be a sign you could benefit from the work of these pioneers.

Consider what the ‘New School’ gurus have to say. Today we have so many great authors and teachers we can also learn from. Alberto Cairo, Elijah Meeks, or Cole Nussbaumer Knaflic — the latter the best-selling author and organizer of the “Storytelling with Data” workshops — all of them educate data visualization students around the world. In October 2019 I attended a workshop with Cole and her staff, where I had the opportunity to test my knowledge with her and colleagues from other companies and industries.

If you have read their blogs, listened to a podcast or attended a seminar, what did you learn? Do you have a sense of what you can apply and test with your audience? Use their advice as the basis of creating your own visualization style.

Enroll in a university program or earn a certification. I am currently attending the Data Visualization certificate program at the University of Washington. We study topics from visualization history and fundamentals to advanced techniques for designing for user consumption. Through exercises, collaboration, and collective feedback, I am building my visualization skills in a structured manner. My favorite part is transferring what I’ve learned to my day-to-day job.

Apply new knowledge

Based on what I have learned in my data visualization journey, here is the result of how I would now construct the previous chart:

Do you think it is easier to read? Do you need less time to understand the data and extract insights?

To be fair, data visualization can be highly subjective. I might prefer one chart over another, or I might format it in a way that suits my preferences or experience. You might have other preferences and disagree with mine. No single prescriptive model exists for everyone in all circumstances. But as with other applied arts and sciences informed by practice, your work is likely to benefit from the guidelines of visualization experts. Their rules and recommendations are based not on mere opinions but on years of research, scientific studies, and trial and error, revealing what works best in communicating information for impact.

Conclusion

Improving visuals can pay dividends, because it can help stakeholders truly understand and act on data presented to them.

Keep in mind that attention of your audience is limited. You typically have only a short amount of time to make your point. This means you must remove barriers to delivering your insights. Do it by conveying a story your audience cares about, influencing and helping them make better decisions because of your work.

Visualizations are key vehicles to achieving this, and if you do them well, your message will be received and acted upon. Don’t take visualizations for granted. Find ways to learn how to keep making them better.