From Numbers to Actionable Insights: Data Visualization Best Practices

Blog07 Dec 22

Great data visualization is like any great picture. It tells a story. It reveals a deeper truth. It delivers a message. And it does all of this in an instant. That’s the goal.

But how do you achieve that goal? What makes data visualization great? It’s a question that our consultants routinely hear from clients looking to improve the business insights that they are gleaning from their data. The first step is understanding the science of visualization and the types of business questions it can answer. The second step is understanding the people who will benefit from access to those answers. The third step is understanding the best way to relay those answers to those users.

In other words, the best visualizations combine a keen understanding of the data, the technical possibilities, the business practices, and the audience needs to provide direct answers to specific queries in a digestible, user-friendly format that encourages adoption.

Understanding the Power of Visualization

At their core, data visualization is about patterns, and it can help to identify them, explain, and describe them. By transforming numbers and groups of numbers into physical points on an axis, visualizations can help users understand data as they exist in time and space. They are a particularly useful tool for identifying and expressing relationships that exist within a set of numbers. They do this by using visual elements such as size, shape, color, and brightness to represent statistical measures like quantity, quality, change and density in a way that appeals to our brain’s ability to recognize patterns.

A clear visualization requires a clear purpose. Scatter plots are a great tool for data exploration, allowing users to see the way in which various data points are correlated. For everyday reporting purposes, elements like histograms, line graphs and waterfall charts  can help users understand and contextualize key performance indicators at a glance. In agenda-driven settings such as presentations, visualizations that use size and colors to compare and contrast can be helpful. Tools like Power BI offer a wide range of these options. The challenge is identifying which option is best suited for the task at hand.

The best visualizations reduce complex sets of variables into simple form. But don’t let the shapes and colors fool you. Getting from Point A to Point B requires an expert-level understanding of statistical and geometric principles. Today’s tools make it easy to create attractive visualizations. The key is making sure that those visualizations are telling an accurate story. Flawed data or flawed calculations can lead to flawed visualizations and flawed decisions. It can also lead to an environment where users lose trust in the reliability of what they are seeing.

Understanding your Audience

The best visualizations and dashboards are those that people actually use. User adoption is a big point of emphasis whenever Thorogood begins work on project with one of our clients. The only way a company will see a return on its investment is if its employees willingly incorporate the new tool into their workflows. Accuracy is only one part of the equation. A dashboard also needs to incorporate these four fundamentals to connect with users.

  • Usability: The last thing a user wants is to feel lost or overwhelmed when attempting to navigate and digest a dashboard. The elements on a screen and between screens need to be organized in a concise, intuitive manner that creates a user-friendly flow. Achieving this requires an understanding of how specific users will be using the tool. The same data set can generate different dashboards for different users. While executives might prefer a high-level snapshot of data, analysts might prefer a deeper level of detail.
  • Performance: One of the surest ways to discourage adoption is to make a dashboard that consumes too much of a user’s time. The speed of page loads and data refreshes are critical considerations in this regard. A dashboard is only as good as its underlying solution architecture.
  • Business value: Bottom line, users will gravitate toward the tools that offer them the clearest value. The central focus of any report needs to be a specific business objective. In short, it needs to answer a question that its users are asking.
  • Communication:  The whole point of visualizations is to make patterns easy to identify and analyze. The visualizations themselves need to give users a similar sense of ease.

Best Practices

The best way to accomplish all of the above is to make every visualization as simple, concise, and purpose-driven as possible. Visualizations are not an ideal format to explore complexity and nuance. They are a place for main ideas. Each visualization should prioritize its main point, and avoid anything that might distract from a user’s understanding of that point.

  • Understanding Axes and Labels: The goal of any visualization is to relay maximum information in minimal form. Any unnecessary labeling can make for a clunky visual experience that distracts from the important information. One common design flaw is attaching labels to axes that are already self-explanatory. Say you create bar chart comparing the total number of users in the United States, India and Brazil. Do you really need to label the Y Axis, “Countries”?
  • Minimize repetition: Consider a bar chart where the Y Axis measures current month sales in increments of 1,000 euros. If you were to label each interval between 0.00 and 10,000.00 with the full number, you would end up cluttering your screen with zeroes, commas, and euro signs. Instead, you could simply label the intervals 0-10 and title the chart, “Current Month Sales (In Thousands of Euros)”
  • Use color wisely: Color is a great visual element when it is used to relay information, like positive or negative gain. Used strictly as a design element, it can distract viewers by drawing their eye away from the intended focus of the visualization, or by causing them to wonder, “What does that color signify?” When a color does signify something important, it is crucial to keep that consistent across relevant visualizations.
  • Understand how the brain processes information: This is particularly true when deciding how to position elements of a visualization. In the Western Hemisphere, at least, our brains tend to look at things in a left-to-right and top-to-bottom fashion. In a quadrant, the eyes move in the shape of a ‘Z’: from the top left corner to the top right corner, to the bottom left corner, to the bottom right corner.

The ideal visualizations may look simple, but achieving that simplicity takes a great deal of work and careful consideration. Feel free to contact Ben White or Sarah Diehl for a deeper look at the incredible business value that visualization can create for enterprise firms.

Find out more

Contact Sarah Diehl. Sarah is a Data & AI Consultant at Thorogood based in the US.

Contact Ben White. Ben is a Data & AI Consultant and Head of Brazil Operations based in São Paulo, Brazil.