Avoiding Common Data Visualisation Mistakes: A Checklist
Data visualisation is a powerful tool for understanding and communicating insights. However, poorly designed visualisations can be misleading, confusing, or simply ineffective. This checklist outlines common mistakes to avoid, helping you create clear, accurate, and impactful charts and graphs. Before you publish your next visualisation, run through these points.
1. Choosing Appropriate Scales
One of the most common and impactful mistakes in data visualisation is using inappropriate scales. This can distort the data and lead to incorrect interpretations.
Truncated Axes
Mistake: Starting the y-axis at a value other than zero when representing quantities. This exaggerates differences and can create a false impression of change.
Why it matters: Our brains naturally interpret the height of bars or the slope of lines as proportional to the values they represent. Truncating the axis distorts this perception.
Example: Imagine a graph showing sales figures for two products. If the y-axis starts at 90% of the maximum value, a small difference in sales might appear as a dramatic increase. Always consider what Charting offers in terms of ensuring accurate data representation.
Solution: Generally, start the y-axis at zero. If the data range is very narrow and starting at zero obscures meaningful variations, consider using a more appropriate chart type or explicitly indicating the truncated axis with a break symbol.
Inconsistent Scales
Mistake: Using different scales for similar data across multiple charts. This makes it difficult to compare the data and can lead to misinterpretations.
Why it matters: Consistent scales allow viewers to quickly and accurately compare data across different visualisations. Inconsistent scales force viewers to mentally adjust, increasing cognitive load and the risk of errors.
Example: If you're comparing sales performance across different regions, ensure that the y-axis (sales) uses the same scale for each region's chart.
Solution: Standardise scales across all relevant visualisations. If different scales are necessary, clearly label them and provide context to explain the differences.
Logarithmic Scales
Mistake: Using logarithmic scales without proper explanation or justification. Logarithmic scales are useful for displaying data with a wide range of values, but they can be confusing for viewers unfamiliar with them.
Why it matters: Logarithmic scales compress large values and expand small values, making it easier to visualise exponential growth or decay. However, they can be misleading if not properly explained, as equal distances on the axis represent multiplicative changes, not additive changes.
Example: Visualising population growth over a long period might benefit from a logarithmic scale. However, it's crucial to clearly label the axis and explain that it represents logarithmic values.
Solution: Only use logarithmic scales when necessary and always provide a clear explanation of what they represent. Consider whether a standard linear scale with appropriate transformations (e.g., square root) might be more accessible.
2. Avoiding Chartjunk
"Chartjunk" refers to unnecessary visual elements that clutter a chart and distract from the data. Minimising chartjunk improves clarity and enhances the viewer's understanding.
Unnecessary Decorations
Mistake: Adding decorative elements that don't contribute to the data representation, such as 3D effects, excessive gridlines, or irrelevant images.
Why it matters: These elements distract from the data and can make it harder to interpret the information. They increase cognitive load without providing any additional value.
Example: Using 3D bar charts when a simple 2D bar chart would suffice. The 3D effect often distorts the perception of bar heights and adds unnecessary visual complexity. You can learn more about Charting and our approach to clean design.
Solution: Remove any visual elements that don't directly support the data representation. Focus on clarity and simplicity.
Excessive Labelling
Mistake: Over-labelling the chart with redundant information or using labels that are too small to read.
Why it matters: Too many labels can clutter the chart and make it difficult to focus on the key data points. Labels that are too small are simply useless.
Example: Labelling every single data point on a line chart when only a few key points are important. Or using a font size that is too small for the labels to be legible.
Solution: Use labels sparingly and strategically. Focus on labelling key data points and use clear, legible fonts. Consider using tooltips or interactive elements to provide additional information on demand.
Overuse of Colour
Mistake: Using too many colours or using colours that are visually distracting or difficult to distinguish.
Why it matters: Too many colours can overwhelm the viewer and make it difficult to identify patterns. Poor colour choices can also make the chart inaccessible to people with colour vision deficiencies.
Example: Using a rainbow colour scale for sequential data, which can create false patterns and is not accessible to people with colour blindness.
Solution: Use a limited colour palette with colours that are visually distinct and appropriate for the type of data being represented. Consider using colourblind-friendly palettes. Use colour to highlight key data points or categories.
3. Using Consistent Formatting
Consistency in formatting is crucial for creating professional and easily understandable visualisations. Inconsistent formatting can be distracting and can undermine the credibility of the data.
Font Styles and Sizes
Mistake: Using different font styles and sizes throughout the chart.
Why it matters: Inconsistent fonts create a disjointed appearance and make the chart look unprofessional. They can also make it harder to read the labels and annotations.
Solution: Choose a consistent font style and size for all text elements in the chart. Use different font weights (e.g., bold) to emphasise key information.
Colour Schemes
Mistake: Using different colour schemes across multiple charts that represent similar data.
Why it matters: Inconsistent colour schemes can make it difficult to compare data across different visualisations. Viewers have to mentally re-associate colours with categories, increasing cognitive load.
Solution: Use a consistent colour scheme across all related charts. This helps viewers quickly identify and compare data across different visualisations.
Axis Labels and Titles
Mistake: Using inconsistent axis labels and titles.
Why it matters: Inconsistent labels and titles can be confusing and make it difficult to understand the chart's purpose and the data being represented.
Solution: Use clear, concise, and consistent axis labels and titles. Ensure that the units of measurement are clearly indicated.
4. Double-Checking Your Data
This seems obvious, but errors in the underlying data are a surprisingly common cause of misleading visualisations. Garbage in, garbage out!
Data Accuracy
Mistake: Using inaccurate or incomplete data.
Why it matters: Inaccurate data will lead to incorrect conclusions and misleading visualisations. Incomplete data can also distort the results and create a false impression of the underlying trends.
Solution: Always double-check your data for accuracy and completeness. Verify the data against the original source and correct any errors. Handle missing data appropriately, either by excluding it or by imputing values using appropriate statistical methods. If you have frequently asked questions about data handling, consult reliable resources.
Data Transformations
Mistake: Applying incorrect or inappropriate data transformations.
Why it matters: Incorrect transformations can distort the data and lead to misleading visualisations. Inappropriate transformations can obscure meaningful patterns or create false patterns.
Solution: Carefully consider the appropriate data transformations for your data. Ensure that the transformations are applied correctly and that the results are properly interpreted.
Outliers
Mistake: Failing to identify and address outliers.
Why it matters: Outliers can significantly distort the appearance of a chart and can lead to incorrect conclusions. They can skew the scales and obscure meaningful patterns in the rest of the data.
Solution: Identify outliers and consider whether they should be excluded from the analysis or treated differently. If outliers are included, consider using a logarithmic scale or other techniques to reduce their impact on the visualisation.
5. Seeking Feedback and Iterating
Data visualisation is an iterative process. It's important to seek feedback from others and refine your visualisations based on their input.
Peer Review
Mistake: Not seeking feedback from others before publishing or presenting your visualisations.
Why it matters: Others may spot errors or inconsistencies that you have missed. They can also provide valuable insights into how to improve the clarity and effectiveness of your visualisations.
Solution: Share your visualisations with colleagues or peers and ask for their feedback. Be open to constructive criticism and use their input to improve your work.
User Testing
Mistake: Not testing your visualisations with the intended audience.
Why it matters: The intended audience may have different levels of expertise or different perspectives on the data. Testing your visualisations with the intended audience can help you identify any potential misunderstandings or areas for improvement.
Solution: Conduct user testing with members of your intended audience. Ask them to interpret the visualisations and provide feedback on their clarity and effectiveness. Use their feedback to refine your visualisations and ensure that they are easily understood by the target audience.
By following this checklist, you can avoid common data visualisation mistakes and create clear, accurate, and impactful charts and graphs that effectively communicate your insights. Remember to always prioritise clarity, accuracy, and accessibility in your visualisations. Consider our services if you need help refining your data visualisation skills.