Doing data visualizations correctly takes careful consideration. Incorrectly visualizing something can be misleading, embarassing, and even damaging to reputations. In order to do it correctly, it can often be useful to think about the visualization from several different angles before settling on the final version. Looking at good examples of data visualization is certainly a great way to learn, but equal value can be found in examining visualizations that didn’t work so well. I recently started a Tumblr for collecting examples of bad visualizations. The examples are often funny, but #WTFViz is not intended solely to be humorous. The examples are also there as educational material, showing what not to do. The data going into the visualization is the best place to start when selecting what visualization to use. There are subtle but complex concepts contained in data, and those need to be reflected accurately in the end visualization. For example, sometimes percentage data has part to whole relationships, and other times it represents overlapping sets. Sometimes people get this wrong, though, and create things like the man who is 243% Baby Boomer. Another thing that comes from the data is whether or not values are categorical or continuous. Line charts are great for continuous data, while bar charts are good at representing categorical dimensions. Ignoring what these charts are good at can produce things like this Line/Bar Chart Amalgam.

Honestly, it’s wild how often even professionals fall for these basic pitfalls. Maybe the pressure to impress with novelty gets in the way of common sense, or maybe people just don’t have time to sit and question their choices every step along the way. I’ve definitely seen cases in 2025 where a flashy animation or some clever effect just distracts from the actual message—like, you end up staring at the spinning chart, but can’t remember what it was about two minutes later. What’s left is a vague sense of confusion and, occasionally, an email chain asking, “Wait, how is this supposed to help us?” It’s pretty clear: no amount of polish can rescue a fundamentally flawed visualization.

That’s not to say experimentation is bad, because sometimes rules do need breaking. But it really does help to circle back and ask real people if your graphics make sense. You’ll be surprised (or maybe not) at how many “perfect” designs don’t survive five minutes of honest feedback. Sometimes I’ll run a draft by a friend who’s not deep in the data world, and if they can’t explain it to me afterwards, back to the drawing board it goes. At this point, experience has shown it’s a lot less stressful to catch issues early, rather than watching your chart become a cautionary tale on some blog.

Another common mistake is making things 3D for no reason. 3D charts almost never add meaningful information, and usually obscure the data. One particularly bad example is this Spiral Staircase Chart. Clearly labeling visualizations, and drawing diagrams that actually clarify things is also important. After all, the point of drawing it out is to help you communicate it to someone, not to confuse the heck out of them. Venn diagrams also have very definite relationships that they represent. Drawing things that mimic the Euler diagram format without the data to match the format is a great way to obfuscate your data. Unfortunately, the supply for WTFViz.net shows no signs of slowing, maybe because of the stupid things designers do when they are trying to be creative. Regardless of the reason, keep your eye out for bad visualizations that you can turn into learning experiences. Drew Skau is Visualization Architect at Visual.ly and a PhD Computer Science Visualization student at UNCC with an undergraduate degree in Architecture. You can follow him on twitter @SeeingStructure

}}