Go To Search
Click to Home
Data Visualization
By Nicole Dewees

When trying to edit my audit reports, I often ask these four questions:

1. Is this necessary?
2. Is this easily understood?
3. Does this convey what I intended?
4. Would this be more powerful as a graph?

Using these four questions, I will go through the steps of improving the data visualizations in your audit reports.


This first question helps me cut unnecessary charts. Researchers have found that people’s eyes are drawn towards images, including graphs and tables. In fact, people will often read about 20 percent of the text, but look at nearly 100 percent of the visuals. Thus, if there’s a graph on your page, it may be the only thing the reader looks at on that page.

Using this question, I looked back at my first audit report and I found a graph that I should have deleted. This audit was about the misuse and unnecessary spending on county-owned mobile devices. The graph below adds nothing to this narrative. As a new auditor, I fell into the trap of, “If I have data, I should graph it.” This graph likely drew people’s attention away from the text on the page and did not contribute to the point we were trying to make.



To answer this question, it is helpful to turn to two experts in the data visualization field: Edward Tufte and Stephanie Evergreen.

Here are some of my favorite quotes from Dr. Tufte:

  • “Forego chartjunk.”
  • “It is not about you anymore. It’s about your audience’s relationship with your content.”

Dr. Evergreen has the following concepts that help me with data visualization.

  • What story are you trying to tell?
  • Directly label charts
  • Highlight what is important

I’ll use some real world examples of graphs and use these concepts to show how they can be improved.

The graph below came from a financial report from a government agency. The first piece I will point out is what Edward Tufte calls “chartjunk,” which is anything that prevents the message from being clear. The first element I would remove is the 3D effect. This effect makes things in the foreground appear to be larger than elements in the background. Beginning Working Capital appears to be only slightly smaller than Intergovernmental in the pie chart. In addition, the categories are not sorted from largest to smallest, making it difficult to quickly know how budget types compare with each other.


I fixed these two problems in the graph below. Now you can see that Beginning Working Capital appears much smaller than Intergovernmental and you can very quickly ascertain which categories are largest and smallest.


For the next improvement, I will use Dr. Tufte’s concept of, “It is not about you anymore. It’s about your audience’s relationship with your content.” In this scenario, I am interested in all of the categories, but my audience is not. Thus, I should only focus on what information the reader needs. I can do this by combining categories, as shown below.


Next, I would use Dr. Evergreen’s concept of, “What story are you trying to tell?” Rather than hoping that the audience will have the same takeaway as me, I need to make what I want them to know abundantly clear. To do this, I need to add a descriptive title, rather than the typical title which would say something vague, such as, “Government Revenues.” To make the message clearer, I used a bright contrasting color for Intergovernmental and used that same color in the title. This is based on Dr. Evergreen’s concept of highlighting what is important.



The next question I like to ask myself is, “does this convey what I intended?” I received the following chart from someone who was trying to communicate that their clients tend to be school-aged. I do not think this chart successfully communicated this message, so let’s see what steps should be taken to improve it.


First of all, this chart is sorted from the largest to the smallest group. That makes sense for many categories, but not numerical data, such as age. In addition, there are simply too many categories for a pie chart. A bar chart, sorted by age, makes it much easier to read.


Next, I would want to make the message clearer by having a more descriptive title and using color to highlight what is important.


Finally, I would remove the “chartjunk” and take out the lines and axis. This focuses the message on what we want, rather than it being too visually overwhelming. I would also use Dr. Evergreen’s concept of directly labeling the bars, rather than making the reader try to figure out the values by looking at the axis.



The final step I take in improving my audit reports is scanning the text for something important. Since people are much more likely to look at an image, I should try to visualize anything that could get lost in a paragraph. Some data might seem like it would be difficult to create a compelling graph, such as a simple percent. However, the Portland Auditor’s Office found a clever way to highlight an important percentage in this graphic. Rather than let this significant number get lost in the text, they used easy-to-understand graphics to demonstrate how few short-term rentals are permitted.


Before publishing your reports, try asking yourself these four questions. When in doubt, a fresh pair of eyes from a colleague who is not working on the report can be helpful in answering these questions.


Nicole Dewees is a Principal Management Auditor in the Multnomah County Auditor's Office. She has a Master's degree in Business Administration and is a Certified Internal Auditor. Prior to joining Multnomah County, Nicole worked as a management analyst for the Washington State Employment Security Department and as a researcher for the Washington Small Business Development Center. She has been awarded Tableau’s Viz of the Day on two occasions and has co-authored two Knighton Award winning audits.