Bad Data Visualization: 9 Examples to Learn From

Imagine you had a big block of marble. In the hands of Michelangelo, that block would turn into a beautiful sculpture with stunning attention to detail. In the hands of the person writing this article, that block would turn into millions of pieces of broken marble.

The same applies to data points. Starting off with the same data, you could have two different results: a dashboard with insightful graphics or something not even Bill Gates could decipher.

In other words, a bad data visualization. Today, we’ll show you what that is and how to avoid it, as well as provide some examples you can learn from.

What is a bad data visualization?

A bad data visualization is the visual representation of data that misinforms the reader and leads them to the wrong conclusions. Instead of making data-driven decisions from relevant data, the reader is left confused.

Even with proper underlying data analysis and clean datasets, using the wrong data visualization techniques can lead to charts and graphs that simply make no sense.

When we talk about misleading data visualizations, it usually has at least one of the following properties:

• Trying to cram in too much data in one graph
• Using the wrong type of chart for your intentions
• Using colors in the wrong way
• Lack of context
• Incorrect data handling before the visualization
• Using the wrong data visualization tools (e.g. Excel for product analytics)
• Lack of storytelling
• Not understanding the target audience

These are just some of the most common mistakes that can happen to your data presentation. But let’s take a look at some bad examples and see what you can learn from them.

A map with poor use of colors

Here is an example of a map, showing the number of US Americans with ancestry from different European countries.

It does not take long to figure out what is wrong with this visualization. According to data visualization best practices, colors should be used strategically to amplify the main message. Colors are not labels and should not replace them.

In this case, the colors make no sense whatsoever as the same color is used for different values on the map. Here is what a good use of colors looks like in a map chart:

This chart is more effective because of the good use of the color scale. The lightest colors have the lowest value, and the darker the color gets, the higher the value. Using different shades of the same color to show values on a scale is a good use of colors for data visualization.

Improper use of the y-axis

Political campaigns have a tendency to twist the truth and this bar chart is a great example of that.

The two results are very close to each other (46% vs 47%) but the wrong scale of the y-axis makes it seem like the bar on the right is significantly bigger than the one on the left.

In this case, the difference is easy to see with a human eye. However, the more data you add and the more complex your charts become, the harder it becomes to spot this kind of bad data visualization.

Here is another example where the x-axis is stable but the y-axis is questionable, to say the least:

Choosing the wrong type of visualization

At first glance, this seems like a regular infographic. But if you want to draw some conclusions from it, you’ll soon be left scratching your head.

The graph should display the different ages at which young people leave their parents’ homes in European countries.

The main reason why this is a poor data visualization example is because of the chart type. These balls or clouds tell the reader nothing and in fact, confuse them even more. Ideally, this should be a map, a line chart, or something along those lines.

The wrong use of legend (and colors)

This chart shows the most used messaging apps broken down by country. If you want to break into a new market, this could help with decision-making, but there’s only one problem. 

You don’t have to be colorblind to realize that the colors are too adjacent to each other to get anything meaningful out of the visualization. Also, the legend is practically useless as it does nothing but lead to misinterpretation of data.

Driving misinformation by putting data out of context

This grim data visualization shows the number of serial killers per country in the world, ranked on a bar chart. There are two problems with it.

The first one is that the values in the bar graph show the absolute number of serial killers per country. In other words, the number of serial killers are not shown in proportion to the inhabitants of the country. 

If you compare a country of +300M inhabitants (like the US) with a country of 67M inhabitants (like the UK), you’re comparing apples to oranges. The end result is a huge horizontal bar for the United States, which is just a poor example of data analytics.

The second issue is the map below the chart simply makes no sense in relation to the chart data. It’s unclear what the dots represent, and there are no tooltips or info panels to provide context.

Pie charts and donut charts: tasty but not very effective

We’ve talked about pie charts and donut charts numerous times on the Luzmo blog. In business intelligence applications and elsewhere, they are hard to read and interpret. This donut chart shows the number of outbound airport passengers, comparing the values from 2019 and 2024.

Even if there was a legend below, comparing the data sets and making logical conclusions is incredibly difficult.

Here is one more chart that tries to simplify the underlying data but in the end, makes things even more complicated and confusing:

If you’re looking for a better alternative to donut and pie charts, we suggest using bar charts.

A scatter plot that tells nothing

This is an example of a scatter plot comparing different digital strategy consulting services.

In this case, the problem is that when everyone is a leader, no one is a leader. In other words, the problem is in the choice of the visualization type.

Data visualization best practices to follow in 2024

We’ve covered this topic many times before, so let’s give you a quick recap of the most important data visualization best practices to follow. If the underlying data science is done right, just follow these tips to make your dashboards, spreadsheets and reports easier to understand and read.

1. Create visualizations with a specific target audience in mind
2. Use your colors carefully - up to 6-8 colors per visualization, and use them to reinforce a message, not tell a story
3. Avoid excessive detail and cluttering up your visualizations
4. Use the right type of visualization for the purpose you have in mind
5. Highlight only the most important information in the visualization
6. Keep all your visualizations on one screen
7. Use legends and tooltips to explain what is happening in the visualization
8. Provide context for your data whenever possible
9. Use the layout of the visualization to tell a story
10. Structure the visualization so that the most important information is displayed first

Wrapping up

Bad data visualizations lead to costly errors. Your target audience won’t be able to read and understand them and make data-backed decisions. What seems like a small confusion can lead to massive consequences for your business.

If you want to create a dashboard for your SaaS product, we can help you make sure you choose the right type of visualization. In Luzmo, you can use dashboard templates to get the right kind of chart or graph for your needs. 

Sign up today and build your next product dashboard for free!