Table of contents:
Why Not All Data Visualizations are the Same
What is a Bad Data Visualization?
Data visualizations are one of the most compelling and effective ways to understand the world around us. This is because data visualizations tell stories, creatively presenting a dataset to an audience to inform them about a given issue or phenomenon. While visualizations have become increasingly complex, informative, and sophisticated, they sometimes go a little too far…providing too much information or becoming too creative for their own good.
This renders visualizations misleading or confusing rather than helpful to their intended audience, and overly complicated rather than easy to understand—defeating the very purpose of visualization to begin with.
There are several ways by this problem can occur:
- Poor color choices,
- Disproportionate sizing of visual components,
- The presence of extraneous or redundant data, or
- Arranging the data in a way that simply doesn’t make any sense.
In this article, we’ll take a look at examples of bad data visualizations. We’ll break down where the graphic went wrong—but we’ll also explain how to improve it, so it can better present the story it’s trying to tell.
1. Unclear Data Direction
The Problem 🤔
In this graphic, we see the average ages at which European young people leave home: Nordic countries the youngest, Western and Central Europe in the mid-20s, and Eastern Europe and the Mediterranean countries in the late 20s (or even later!). Bubbles represent countries, each including the national average age of “flying the nest.” Bubble coloration is dependent on this average age.
Overall, this isn’t a bad concept. But here’s the problem: the bubbles are not arranged in any meaningful sense. The red countries seem to cluster somewhat at the bottom, and blue countries at the top—but that’s it. This might be a bad take on a map of Europe, but we can't be sure.
Bizarrely, the bubble indicating the average EU age of leaving home is barely identifiable, located on the far left of the graphic and indistinguishable in size and format from any of the country bubbles. In short, the missing key here is organization.
The Solution đź’ˇ
A more effective means to organize this data would be to construct it in color-coded layers, with the bulk of yellow countries falling in the middle, and tapering off toward the blue and red categories above and below.
The EU bubble could either be placed off to the side or in the center of the yellow layer. Regardless of its placement, it should have some distinguishing feature—such as larger size or bolded text—to indicate its importance.
These fixes would allow readers to not only know the differences between “flying the nest” ages by country, but also understand how these data points fit together, allowing them to draw correlations and conclusions about this social phenomenon.
How Plecto helps solve unclear data direction
With predefined visualizations like bar charts, line charts, speedometers, and more, Plecto guides users toward common formats that have clear axes, hierarchy, and organization baked in their design.
2. Too Many Variables
The Problem 🤔
This chart is just a classic case of TMI (too much information)! The data here are measured on not just a typical two, but three axes: number, month, and type of fruit. While we can appreciate the multidimensionality of this chart and its effort to communicate as much data as possible, it’s just too much to understand at first glance.
And this leads to some serious problems when it comes to interpreting this graphic. It’s difficult to ascertain the number of lemons across all months, for example, and impossible to see the number of bananas in February and March.
The reader also can’t accurately tell the number of fruits for any of the bars in the central portion of the graph. If they’re not placed directly against the number grid, the amount of fruit is basically impossible to determine.
The Solution đź’ˇ
The solution to this graph’s problem is to eliminate one of the axes and have the new graph measure only two components instead of three. While this will result in more graphs to convey the same amount of information, they will be far easier to read.
For example, one could track the number of oranges sold across the four months, and make corresponding graphs for apples, lemons, and bananas. Alternatively, one could track the four types of fruit sold in January, and make corresponding graphs for the remaining three months.
How Plecto helps you avoid using too many variables
Allowing users to easily filter, stack, or separate variables into multiple, readable charts, Plecto allows sales and support managers to analyze key metrics by employee, team, time, or other variables on different widgets on the same dashboard. This versatility prevents the impulse to cram too much data into a single, confusing visualization, while still offering powerful insights into performance.
Start your 14-day
free Plecto trial today.
3. Counterintuitive Pie Chart Percentages
The Problem 🤔
The concept for this bespoke pie chart is creative and decidedly unique. But man, we’ve gotta say it—the math isn’t mathing! While every side of a hexagon is equal to one-sixth, or 16.67 percent of the hexagon’s total area, what’s seen here completely upends this basic reality.
Somehow, the fructose section (38.5 percent) takes up vastly more space than the glucose section (31.0 percent), which in turn is only slightly larger than the water section (17.1 percent). And the 0.5 percent sliver of minerals, vitamins, and enzymes is far too large for its stated percentage.
The Solution đź’ˇ
It’s fine to put a creative spin on a typical, “boring” pie chart—but this shouldn’t come at the expense of statistical accuracy, or just plain common sense. Thankfully, the fix for this botched honeycomb chart is a simple one: restructure the graph sections so that they are in proportion to their stated values.
How Plecto helps avoid disproportionate pie chart segments
Plecto ensures statistical accuracy, with donut charts that are automatically calculated to render segment size perfectly proportional to its stated value, guaranteeing a visual representation that is always truthful to the data. Plus, did we mention, they look great?
4. Disproportionate Pictogram Sizing
The Problem 🤔
This is another case of incorrect proportions, but of a different type. In this case, two bar charts indicate the number of graduates in A and B cohorts: one graduate in A, and three in B. But instead of using bars, the charts use pictograms. The left-hand graph uses discrete units to count the number of graduates; meanwhile, the right-hand graph uses a single large pictogram to indicate the same quantity.
However, this is where the problem for the right-hand chart arises. Rather than simply indicating quantity, using a scalable pictogram introduces another element of dimensionality—not just of number, but of size.
This creates situation where, in order to keep the pictogram proportionate while hitting the “3” mark, it is scaled to a size that leads the audience to believe the difference between A and B is much greater than it really is. Indeed, the B graduate doesn’t appear merely three times larger than the A graduate, but rather six or seven times larger!
The Solution đź’ˇ
The solution to this poor-quality graphic (on the right) is found on the much better graphic on the left. Here, three pictograms of equal size are stacked on each other, proportionally (and correctly) indicating the difference between 1 and 3.
When thinking of how to display data, one has to remember what exactly is being represented: difference in proportion (the relationship between parts of a whole), or difference in numerical quantity (comparing values or quantities across categories or time).
If you want to see a proportional representation, use visualizations such as pie (donut) charts, stacked bar charts, or tree maps. But if you’re looking to simply compare dataset sizes, use discrete units—or a simple bar—to represent them realistically.
How Plecto helps you avoid disproportionate unit visualization
Forget scaling issues when you can automate visualization. Plecto uses fixed sizing for unit comparison, preventing scaling errors. For quantity comparisons, standard bar and column charts are used by default to accurately show discrete unit differences. Not only is this a time-saver, but it looks professional and shows performance exactly as it stands.
5. Distorted Pie Chart View
The Problem 🤔
As an example of proportionate rather than numerical representation, the two pie (donut) charts below display proportionate shares of “cause of death.” Looking at the chart on the right, we’re led to believe that section B is the largest section—or at least of equal size to section A. But looking at the chart to the left, we see this isn’t the case, and section A is quite clearly the largest. What gives?
The 3D pie chart view on the right distorts the graph, with section B tilted toward the viewer and thereby making this section appear far larger than it really is. In short, it’s a trick of the eye—one that, while not presenting false information outright, creates a misleading perception of reality.
The Solution đź’ˇ
The solution here is really simple: avoid using 3D pie charts, which invariably distort the data you’re presenting, and use 2D charts instead. Doing so is more honest, and also ensures you’re getting the most reliable and accurate data possible. This is especially important because so much of what counts as “good” data visualization relies on first glance impressions to establish not only readability, but truthfulness.
How Plecto helps you avoid distorting data
Plecto exclusively uses 2D visualizations for core charts like donut, line, and bar charts, eliminating the potential for distortion caused by 3D perspectives, guaranteeing the visual truthfulness of your data at first glance.
6. Where to Begin…?
The Problem 🤔
For the grand finale, a visualization with not just one or two, but several issues! This graphic shows how members of the “Baby Boomer” generational cohort describe themselves, with each descriptor indicated by a different color-coded percentage on a humanoid figure.
This graphic is ultimately a “creative” pie chart, much like the previous honeycomb example in this list (Number 3). However, while the honeycomb graphic simply didn’t represent proportions correctly, this graphic commits multiple statistical and design sins:
- The percentages listed don’t add up to 100, but rather 243. Here, where no overlap between categories is present, the viewer is led to believe the graphic designer has made a severe statistical error.
- The spaces allotted to each category don’t correspond at all with their given percentages. Quite clearly, the “people-savvy” category (78 percent) does not take up 78 percent of the humanoid!
- It’s also uncertain why the designer chose to use a humanoid and a seemingly random color scheme. What information do these design choices impart to the reader, or what information are they trying to reinforce in the reader’s mind? We simply don’t know.
Overall, this graphic gives a strongly misleading impression about the information it’s trying to convey. When the reader sees the humanoid, they mentally associate this with a totality (such as 100 percent) in the same way as the circle of a pie chart.
But when the reader sees category percentages that exceed 100 and have no visual basis within the graphic, they conclude the numbers are randomly placed and have no real “story” behind them. In other words, the graphic becomes meaningless.
The Solution đź’ˇ
The only solution to fixing this graphic is to scratch it entirely and start afresh—this time, as a bar chart. Transferring this graphic’s data to a bar chart format will allow the intended audience to better understand the relationship between these value categories (“people-savvy,” “creative,” etc.) without giving the audience a false idea that these values are supposed to numerically “fit” together.
If numerical data rather than just percentages can be obtained, an even better solution would be to create a Venn diagram. The bubbles’ overlap would allow the viewer to know more about the relationships between respondents’ values: for example, how many respondents said they were both “leaders” and “willing to learn”? And to give the viewer a quick idea of how many respondents fall in each value category, the bubbles can be sized accordingly.
How Plecto helps simplify data visualization
Plecto's range of 10+ data visualization widget types gives users a range of formats to represent key information, even when there are overlapping variables or individuals falling into two categories–such as a Baby Boomer considering themselves both tech- and people-savvy.
Choose from speedometers, tables, donut charts, area charts, or column (bar) charts to visualize performance KPIs in a way that makes the most sense for you. In fact, the key principle of Plecto is to use data in a simple yet motivating way to boost team performance!
Want to learn more? Why not start a 14-day free trial or book a demo to see how great data visualization can be the difference between hitting next quarter's targets?
Want to Avoid These Bad Visualization Mistakes? Try Plecto!
Because data visualizations are ubiquitous, it’s no surprise that there will be both good and bad examples in abundance. This is why it’s important to know the warning signs of a bad data visualization, so that you can ensure the data you interact with is presented in a reliable, valid, and trustworthy way.
With Plecto, you can avoid the pitfalls of poor data visualization and rely on tried-and-true visualization methods, ensuring your organization can access its data transparently and accurately. This clarity will give you the insight you need to make data-based decisions and hit your KPI goals.
Don’t waste your time on substandard data visualizations! Sign up below for a free Plecto demo today.
Start your 14-day
free Plecto trial today.
Q&As
Why is the "Truncated Y-Axis" considered one of the most deceptive practices in data visualization?
A truncated Y-axis occurs when the vertical axis of a chart does not start at zero. This is a bad visualization practice because it artificially exaggerates small differences between data points, making them appear much more significant than they actually are. This is often used to mislead an audience into seeing a massive growth trend or a sharp decline when the actual numerical change is negligible. To maintain integrity, axes should generally start at zero to provide the proper context for the data being presented.
What is "3D chart overload," and how does it distort data?
Answer: 3D chart overload happens when designers add three-dimensional effects to simple charts (like bar or pie charts) purely for aesthetic appeal. This is problematic because the extra dimension adds unnecessary "chart junk" that distorts the viewer’s perspective. In a 3D pie chart, for example, the slices in the foreground appear larger than the slices in the background, even if they represent the same value. This makes it nearly impossible for the human eye to accurately compare the proportions of the data.
How does Plecto’s focus on dashboard design help users avoid the pitfall of data overload?
Data overload occurs when a visualization tries to cram too many metrics or colors into a single view, leaving the audience confused and unable to identify the most important information. Plecto helps users avoid this by providing clean, customizable templates that emphasize clarity and hierarchy. By using Plecto, managers can select only the most relevant KPIs and use features like "Conditional Formatting" to highlight only the data that requires immediate action. This design-first approach ensures that dashboards remain a "glanceable" tool for decision-making rather than a cluttered wall of numbers.
JAMES NIILER
Content Specialist
An in-house content writer and specialist at Plecto, James brings an academic touch and journalistic flair to his marketing copy. Having worked and studied on both sides of the Atlantic, James is a great believer in the importance of communicating across cultures and industries. Catch his work here on the Plecto blog, or as a guest contributor on other B2B websites.
