The Problem With Your Data Visualization (And How to Fix It)




“The secret language of statistics, so appealing in a fact-minded culture, is employed to sensationalize, inflate, confuse, and oversimplify.” ― Darrell Huff, How to Lie with Statistics, 1954

When Darrell Huff wrote those words over a half-century ago, the world was not amidst a global pandemic, the first cell phone would not be sold for another 30 years, and a global communication network was all but inconceivable.  Yet his description of the sensationalization and distortion of facts are very much at home in the instant gratification and social-media-driven world of today.

“Data Visualization” was not a term widely used back then. However, today as we enter a world of constant virtualization, you can’t escape the charts and graphs that Darrell Huff was undoubtedly familiar with when writing, “How to Lie with Statistics.”

Huff wrote his book as a tongue-in-cheek warning for those who might find themselves confused or deceived by data, and today we feel it’s worth revisiting and seeing how they apply to digital products and websites.

Some Data Visualizations Are Great

To be clear, data visualizations are indeed essential for businesses and other organizations to operate. They allow humans to make sense of raw data which fuels conversations, innovation, and strategy.

However, just because you’ve taken raw data and transformed it into something beautiful does not mean that it’s useful. Data visualization is not inherently good. It is only good if it provides value to your user base.

Some Are Not

There are four main scenarios where data visualizations can hinder–not help–your users.

  • You don’t understand the type of data you need to display
  • You understand the data, but not how to display it
  • You’re overwhelming users with too much data or information
  • A user understands the data but not how to apply it

How to Make a Data Visualization That Works

If you start with the four problems above, you can find ways to make a data visualization that works.

  1. Understand the type of data you need. Where are you collecting it from and where are you storing it? Is there data you need that you currently don’t collect? What’s the best and most secure way to gather that data?
  2. Once you have the data, consider the ways you might display and interact with it. Here are some of the most common charts and graphs, a good place to start.
  3. Think about your user. How data-savvy are they? When and where will they be interacting with your data?
  4. What are the conclusions you want users to gather from your data? Spell it out so it’s not a mystery.

Going through this process will get you 90% of the way to a solid data visualization. A few other tips from our UX/UI team:

One visualization = one goal

Make sure your visual is concise. It’s easy to get caught up in the “cool” elements and forget that normal people need to be able to comprehend what you’re showing.

Define your axes

Working in a particular industry, we often forget that outsiders aren’t familiar with the terminology. For example, the difference between users and sessions in Google Analytics. Defining this upfront will ensure that users fully grasp the meaning behind the raw numbers.

Draw conclusions for your audience

Even if your data visualization is straightforward and shows an “obvious” outcome, don’t be afraid to spell out the results too. Some people need you to connect the dots for them to get the whole picture.

Examples of Data Visualization Done Right

Here are two examples of strong data visualizations.

2020 Election Forecast – by FiveThirtyEight

Why This Data Viz Works:

  1. Many data visualizations on one page allow a story to be told. However, each individual component has a single data goal.
  2. There are conclusions drawn from each graph (supported with their little fox mascot)
  3. Clear axes and labels

35 Years Of American Death – by Five ThirtyEight

Why This Data Viz Works:

  1. This data visualization is showing massive amounts of health data but aligns a single map with a single health issue.
  2. There is categorization information further down the page, in case users have clarifying questions.
  3. Conclusions include counties with the highest mortality rates, counties with lowest mortality rates, and trends across all categories.

Today, you’re probably collecting a ton of raw data, but make sure you use it, display it, and draw conclusions in a direct, unbiased, and clear way. Otherwise, your beautiful, expensive chart may be useless.

If you need help strategizing around data and how to use it to fuel growth, contact Devetry. We take a holistic view of your capabilities and create data visualizations that help you achieve your goals.