I analyzed 100 dashboards. Here are the most common data viz errors I saw.

Photo by Campaign Creators on Unsplash

What data viz errors are most common for entry-level data analysts?

To answer this question, I evaluated 100 dashboards (selected randomly) from the Maven Analytics showcase, where data analysts upload data analyses from Tableau, Excel, SQL, Power BI, or other platforms as part of their portfolio.

I wanted to look into:

• Whether there were surprising recurring errors that data analysts make.
• Examples of good portfolios — what did data analysts do to make their dashboards easier to understand, even if the viewer doesn’t have any familiarity with the subject matter?

Read on for the 8 errors I looked for, how often they occurred, and some surprises I encountered along the way.

Error #1: Unnecessary Visuals (58% of dashboards included this error)

To make your dashboard as effective as possible, eliminate visual elements that take up space but don’t increase understanding (this is called “chartjunk.”) It may be tempting to add squiggly lines, icons, colorful inserts, varying fonts — but if your dashboard is too cluttered, it is distracting and harder for the reader to understand the key data points.

Error #2: Lack of Guidance (53% of dashboards)

I used “lack of guidance” to describe dashboards with errors such as:

• not showing trends or context in the data, either in the dashboard or in the written analysis. Is that number good or bad? We can’t tell without context.
• Axis intervals, reports, x-axis, y-axis, pie slices, bars on the bar chart, etc., are not labeled.
• Context is missing: For example, if a dashboard says “50% decrease!” but we’re not sure what the decrease is in, or for what time frame.
• Legend is missing, so we do not know the meaning of the different colors in the visualization.

I was shocked by how many dashboards did not include units of measurement. I also noticed that many dashboards did not call attention to or explain outliers in their written analysis.

Error #3: Unhelpful Titles/Labels (45% of dashboards)

Analysts often used the default Power BI or Tableau labels. This leads to generic labels that will not make sense to the average reader, such as “SUM_of_TotalIDS_adjusted” or “Total_Bft_Orders_by_Day_Name_English.”

To avoid this error, review all labels and titles in your report before publishing, and change the default labels to something that makes sense to the average reader. If you need to include acronyms or business terms, explain their meaning in your written analysis.

Error #4: Inappropriate Chart Type (42% of dashboards)

I used “Inappropriate Chart Type” to describe errors I saw, such as:

• Using treemaps or pie charts to show change over time (a line chart is better for this.)
• Using pie charts with over 5 segments.

To choose an effective visual for your data, follow Maven Analytics’ advice and use bar charts to compare numerical data across categories, line charts to show trends over time, and scatterplots to show the relationship between two variables.

Don’t use fancy new visuals just to make your dashboard look cooler. If bar charts are the most effective way to display your data, by all means, use bar charts!

Error #5: Poor Use of Colors (40% of dashboards)

Examples of poor use of colors include:

• Abrasive colors for visuals and fonts, including neon green and neon yellow. These colors are hard to read and hurt your eyes.
• Colors mean different things in different visuals on the same page, leading to confusion. For example, pink = “female” and blue = “male” on one visual, and then pink = “urban sales” and blue = “rural sales” on another visual.
• Using the same color for multiple slices of a pie chart, so you can’t tell the difference in slice size.

Error #6: Overly Flamboyant Background (22% of dashboards)

If you are analyzing pizza stores, it may be tempting to have a big pepperoni pizza in the background of your dashboard. But this is just another form of distracting “chartjunk” that detracts from the key points in your data. Resist the urge! Have a plain white or neutral background. This will be much easier to read.

Error #7: Spelling/Grammar/Punctuation Errors (20% of dashboards)

Remember to check your dashboard and your written analyses for any spelling, grammar, punctuation, or capitalization errors before publishing! I’d recommend using Grammarly or a similar app to check your writing.

Error #8: Pie Charts with Over 5 Segments (12% of dashboards)

If a pie chart has over 5 segments (or even over 3, I would argue), it is hard to figure out the angle of each of the slices. Instead, if you have over 5 categories to compare, use a bar chart!

Shout-Outs: Some Excellent Dashboards to Review

While doing my analysis of 100 dashboards, I also found some excellent analyses, and want to make sure I give the authors credit!

The best analyses did the following:

• Explained any acronyms or business/industry terms in simple, easy-to-understand language, so I could follow along even if I didn’t have expertise in the area.
• Compared trends before and after a significant event (for example, comparing hospital satisfaction scores before, during, and after COVID.)
• Graphed and called attention to outliers, providing possible explanations (for example, calling attention to states that had a larger decline in hospital satisfaction during COVID.)
• Provided actionable and practical business recommendations.

If you want to see some stellar dashboards, check out the following, and leave them a positive comment!:

• Pandemic Prescription: Are U.S. hospitals losing their recommended dose of trust? (Winner of Maven Analytics challenge) by Jasmin Jusufbegovic
• My Journey to Love — Romance Novels by Leah Warner-Tedaldi
• Apple’s Progress Towards Carbon Neutrality by Brenda Loznik
• MySQL — Maven Fuzzy Factory by Miguel Fuguet

Further reading:

• Maven Analytics 6 Pro Tips for Dashboard Design will help you choose the appropriate visual for your data.
• Chartjunk: How to Avoid Confusing Elements in Your Figures

The contents of external submissions are not necessarily reflective of the opinions or work of Maven Analytics or any of its team members.

We believe in fostering lifelong learning and our intent is to provide a platform for the data community to share their work and seek feedback from the Maven Analytics data fam.

Happy learning!

-Team Maven