As I discuss in the pre-workshop video for my Practical Charts course, many charts don’t do a great job of serving the purpose for which the chart was created in the first place. I think that there are several reasons why this is such a common problem and this post focuses on a big one, which is that there’s a lot of confusion around exactly what skills are needed in order to create truly useful charts.
Specifically, many organizations seem to assume that people with deep expertise in graphic design, data analysis, and/or data visualization software products must be good at creating useful charts. This is often not the case, however, for reasons that I outline below. While elements of these disciplines are required to create useful charts, they’re not, in and of themselves, sufficient. Why not? Let’s have a closer look at each of these three types of expertise to find out.
Graphic designers, graphic artists, and similar roles
Most graphic design training focuses on how to create images that are visually appealing and, yes, this is an important skill. The purpose of most charts isn’t to be visually appealing, however, it’s to be useful. Most of the time, the visual appeal of a chart has little to do with how useful it is, and the skills required to create useful charts aren’t the same as those that are required to create visually appealing graphics. While charts shouldn’t be ugly, once we’ve applied some basic graphic design best practices around color, decluttering, etc., we’re faced with decisions such as choosing the most effective chart type for a given data set and purpose, deciding if we need to include zero in the quantitative scale, deciding whether to show absolute values or percentages, etc. Most graphic designers have had little training on how to make these types of design decisions.
Expert users of Excel, Tableau, Qlik, JMP, etc.
Assuming that expert users of software products such as Excel, Tableau, or Qlik must all be good at creating useful charts is like assuming that expert users of Microsoft Word must all be great writers, or that knowing every feature of PowerPoint makes one a great presenter. Yes, we need to know how to use those products to create useful work, but product knowledge alone isn’t enough. We also need many higher-level skills that aren’t specific to any particular tool, but that are absolutely necessary in order to create useful work. Unfortunately, though, the software training programs that I’ve come across tend to only cover what CAN be done with a given product (e.g., how to create a bar chart, how to create a line chart, etc.), but spend little time covering actual data visualization principles and best practices (e.g., WHEN to use a bar chart vs. a line chart).
Data analysts, data scientists, and similar roles
We’ve all probably endured presentations or reports that were created by people who were clearly brilliant analysts, but who had never been trained on how to communicate their findings to others. This is probably more proof than most of us need to realize that the skills that are required to analyze data are distinct from those that are required to communicate data to others. Unfortunately, however, most data analysis training programs devote little time to teaching analysts how to communicate their findings to others (which I see as a major gap in such programs).
So, what ARE the skills that are needed to create useful charts?
As I see it, elements of the aforementioned skills are required to create useful charts. For example, we need a certain amount of data literacy (a data analysis skill), some knowledge of the how the human visual system works (a graphic design skill), and, obviously, we have to be basically competent with at least one data visualization software product. This combination of skills, however, is still not enough. We also need other skills that aren’t normally taught in any of those disciplines, and which my friend and colleague Stephen Few has called “data communication skills”. These include skills such as how to choose the most effective chart type for a given situation, how to make the main take-away of a chart visually obvious, how to choose quantitative scales effectively, and how to decide what data needs to be included in a chart in the first place, among many other skills.
In my Practical Charts course, I visualize the relationships among these skills like this:
Unfortunately, though, most organizations don’t seem to understand these relationships and distinctions. This can be seen in their job postings for data visualization roles, which usually read like postings for graphic designers, data analysts, or software power users (the latter being most common). It’s hard to blame them, though, since “data communication skills” is the best term we have to describe the skills that these organizations are actually looking for, but few people understand what that term means.
By the way, if you’re interested in attending my Practical Charts or Practical Dashboards course, here’s a list of my upcoming open-registration workshops.