Stop Making Big Mistakes in Data Visualization

What Makes a Bad Visualization?

It is critical that the data you visually represent is accurate, insightful, not misleading, information heavy or put simply, dull. The impact of your data visualization is compromised when unfavourable aesthetic choices are made therefore, it is important that designers create attractive visualizations and represent data with integrity.

Although not necessarily the result of malicious intent, viewers can be misled by even innocent mistakes in confusing visualizations.

Here are the 10 most common data visualization mistakes and how to avoid them:

1. Information Overload

Choosing what information to include vs. what to leave out in the pursuit of clear communication is a constant design dilemma. Data visualization is no exception, especially when the data is vast and engaging.

When visualizations contain an excessive amount of data, the information becomes overwhelming, and your story or narrative can become difficult to comprehend.

Why avoid?

• Users are unable to comprehend every aspect of the visualization
• There is a lack of clarity on which aspects need our focus
• Your story or narrative is incomprehensible

Tips:

• Identify the most important data for your consumers to understand first, and then present them with only those pieces of information that are essential to conveying your message effectively
• Avoid the temptation to chart all of your thoughts. More data can be communicated with the use of many visuals
• There shouldn't be more than 5 or 6 colours in a single depiction

2. Introducing Bias

Your data visualization's title, label, annotations, and descriptions all work together to communicate the data's meaning to the viewer.

However, if these features tell a tale that is inconsistent with the data, your users will be left feeling confused.

Why avoid?

• Despite having accurate information, if the text is changed in a way that is deceptive to the audience, their interpretations may change as a result

Tips:

• Verify that there is no prejudice in the titles, labels, and descriptions
• Add additional context only when necessary to clarify the content being shown

3. Using an Inappropriate Chart Choice

Selecting the right representation to visualize your data is a crucial stage in the data visualization process. How do you determine which of several possible charts is the best for presenting your data?

A line chart, rather than a bar chart, would be the preferred method of representation if the goal was to demonstrate a change in sales over time, and visualizing the differences across 3 categories could be visualised using a pie chart, but the differences between them would be more obvious using a bar chart.

Need help choosing the right charts? Read this:

8 Popular Chart Types and When to Use Them

Struggling to know which chart type to use? We’ve made it easy for you. Here are the most common chart types and when to use them!

Data-VizDaniel Cox

Why avoid?

• Choosing the wrong kind of chart might throw off readers or even lead them astray

Tips:

• There is no universally best approach to data visualization so take time to learn which features your visualization should represent

4. Confusing Correlations

Correlations across datasets can be visualised to give consumers a more comprehensive grasp of a subject. Overlaying datasets on the same chart is one technique to illustrate connections between them. However, excessive use of overlays makes it challenging for viewers to recognise connections.

Why avoid?

• Correlation does not imply causation
• Correlations can be misleading if they are not truly connected

Tips:

• Multiple, neighbouring visualizations might assist in bringing attention to potential connections. Enable your consumers to evaluate the information while still drawing inferences

5. Highlighting Favourable Data

When talking about data, time is always a factor. In order to support particular storylines, it is possible to focus on specific time periods and display relevant facts.

As an example, consider the practise of "financial modelling," which involves visualising the results of financial operations. Imagine a graph that displays healthy growth rates over a relatively short time frame, giving the impression that a company is doing well. Unfortunately, expanding the view reveals that the company's resurgence was really a temporary blip in the midst of a steep and protracted collapse.

Why avoid?

• Cherry picking data points can obscure crucial information, providing your users with only a partial picture

Tips:

• Evaluate a reduced-scale representation beside a complete representation, such as in the example above
• Aggregate the statistics for the non-zoomed ones together

6. Using 3D Visuals Without Purpose

In the world of data visualization, we can use various charts to display data. Most 3D charts are no longer commonly used to display common data because they have a high risk of misrepresenting data because our human eyes have difficulty interpreting 3D visuals.

Why avoid?

• Data can be distorted by 3D charts

Tips:

• If possible, use a 2D chart instead
• A bubble plot/scatter plot with a colour gradient can be a good option for representing a value that spans three axes

7. Aversion to Standard Visual Design Choices

Human psychology is influenced by visual design elements. Icons, colour schemes, and fonts all have meanings that influence viewer perception. When designers ignore or reject these design choices in favour of creative expression, it rarely ends well.

It is mentally taxing to analyse data visualizations. The brain may not take the time to decipher the reimagined meaning of familiar design elements during the critical moment of cognition.

Why avoid?

• There are numerous approaches to incorporating creative experimentation into data visualization. Do not divert attention away from the data by forcing viewers to reinterpret common visual associations

Tips:

• Colour is an excellent tool for categorising and supporting your main points, and it plays an important role in user decisions
• To make data easier to read, data can be ordered, and various chart sections can be sized

8. Deceiving Colour Usage

Although different colours help the viewer understand data visualizations, too many colours might be overwhelming. The use of only a handful of distinct hues is required.

Why avoid?

• Users will become confused about which values are the most important.
• Users may need more time to absorb the data presented in a visualization if there is an abundance of colour.

Tips:

• You may differentiate between values by using different colours.
• Colours with a high contrast ratio make information seem more important to the human eye.
• One of the quickest and easiest ways to find the right contrast is to use greyscale comparisons of the two colours you're considering

9. Removing Baselines or Scales

The issue with data visualization is that it has the potential to reveal misleading patterns or even nonexistent trends.

Why avoid?

• Users won't fully understand the information, which can lead to misunderstandings
• When people see that the visualization shows wrong information, they will lose faith in its contents

Tips:

• Prioritise making data visualizations with a y-axis that starts at zero.
• If removing the zero makes sense, add a zero-break to show that *If the small changes are really important, it's okay to not start from zero

10. Using Data Visualization Without Purpose

Data visualization isn't a necessity for delivering information.

Sometimes the numbers in your database will speak for themselves. It may not be required to present some values in a data visualization even though they may convey crucial

Why avoid?

• The use of charts and graphs to display information may not always be required.

Tips:

• Information can be communicated through the use of data visualization; in some situations, its usage is sanctioned, whereas in others, a different method might be better suited

Conclusion

It is very important to construct a narrative around your data and make conscious design choices that best reflect it. Consider what your audience will gain from your visualisation when they look at it.

Remember, data visualisations are analytical snapshots that show numbers in a way that the human eye can understand. Be careful not to design data visualisations that create a bias or could be dangerously misinterpreted.