You really don’t have to be a data scientist to understand the power of data visualization. More importantly, you don’t have to be a data scientist to master data visualization.
Mastering this art—particularly in digital marketing, where data sets can be murky, correlations hard to extract, and results difficult to explain—is essential. Data visualization plays a key role in maintaining client relationships and increasing their understanding of your work and results. With the help of Excel, Tableau, or any other data viz tool, you can easily create powerful charts and graphs that clarify the significance of your findings.
But let’s be honest: Choosing the right chart or graph isn’t intuitive.
We’re all taught about box-and-whisker plots, pie graphs, bar charts, etc., in elementary school, but unless you’re using them every day, it’s tough to determine the best way to display data (and this comes from someone with a degree in economics, the graphiest of degrees). Fortunately, some pretty handy rules can ensure you’re using the right graph at the right time to explain the right finding.
Before we get to the nitty gritty, let’s take a look at the downside of a graph that doesn’t fully explain your data and why it’s so important to make your data as clear as possible.
The Difference Between a Good and Bad Graph
A well-designed graph should be easy enough to understand that a child can quickly scan it and understand the insights it was designed to reveal.
The Challenger Disaster of 1986 exemplifies the devastating effect of not displaying vital information clearly. The night before launch, a team of engineers discovered a potentially catastrophic problem with the O-rings used to seal the bottom of the space shuttle’s solid rocket boosters. These O-rings had a tendency to fail in cold temperatures, like those predicted on the morning of the launch, January 28.
Engineers tried to persuade NASA executives that the launch should be delayed until the problem was solved. It was not, and the seven crew members of the space shuttle tragically lost their lives.
So why didn’t NASA cancel the launch? Take a look at the two graphs below—the same ones NASA engineers showed their bosses—and see what your takeaways are. Is it clear that O-rings would fail at temperatures below 30 degrees Fahrenheit?
Now take a look at this graph by statistician and data viz master Edward Tufte:
Even a child could understand the repercussion of launching in chilly weather. The failure rate of O-rings in freezing temperatures is obvious.
In digital marketing, a great graph isn’t going to save a life, but it will aid or reinforce a decision.
Most clients are busy people with businesses to run. They don’t want to devote huge chunks of time to reading a crappy graph. They want readily digestible information and as much clarity as possible, so they can turn your information into action. Remember, the goal is to convey vital information with a quick glimpse at your graph. A reader should understand the key takeaways in moments.
Before we get right into the graphs, let’s review some definitions that you may not have heard in (many) years.
Qualitative vs Quantitative Data
All data is split into two types—qualitative and quantitative. Those two types are split further into two subsets that categorize all data encountered in the wild.
Qualitative data refers to datasets that can be classified or categorized, like colors or satisfaction scores, but not data that can be measured.
Don’t: Use the mean of your data. There is no average of “blue.”
Don’t: Arrange your data with a line chart or histogram.
Nominal Data: Qualitative data that can’t be put in a meaningful order.
Ex. blue, yellow, green
Ordinal Data: Qualitative data that can be put in a meaningful order.
Ex. unsatisfied, satisfied, very satisfied
The best charts for qualitative data are:
- Bar graphs
- Side-by-side bar graphs
- Pie charts
Quantitative data refers to data sets that are measurable, like number of customers, conversion rates, etc.
Do: Use mean, median, mode, and even standard deviation to get the most out of your data set.
Do: Arrange your data using a line chart, bar chart, column chart, line chart, histogram.
Discrete Data: Quantitative data with distinct values or observations. This is data that can be counted.
Ex. The number of students in a classroom, the number of sides on a die
Continuous Data: Quantitative data with a value that falls within a finite or infinite interval.
Ex. conversion rates, visits, pageviews, bounce rate
The best charts for quantitative data are:
- Line charts
- Bar charts
- Column charts
Picking a Chart
Now that we’ve sorted out which graphs are appropriate for which type of data, there are a few questions to ask before you start plotting your charts.
These five questions are an opportunity to think about what you want to accomplish with your data and evaluate the best way to demonstrate that information. From there, you can narrow down which charts to use and which to avoid.
It’s great to know exactly what we’re trying to accomplish, and we’ve definitely narrowed down our options, but the question remains: “Which is the right graph to use?”
The only way to make that decision is to know the pros and cons of each chart. To show you, we’re gonna use a chart!
These are not the only charts or graphs a digital marketer should use, but they are some of the most familiar to a working professional. All are easy to understand and easy to create within Excel or Tableau (a topic for another time).
There’s a whole world of weird graphs out there too that have been used to represent all sorts of unique insights from a variety of different data sources. Websites like FiveThirtyEight, Flowing Data, and Information Is Beautiful host a tremendous amount of weird graphs that display some awesome information.
Have you been using any? We’d love to hear from you.