Best Practices

Follow the principles, process and usability guidelines below as you consider how to create a data visualization.

Principles

Make It Clear

There are many ways to visualize any set of data. Always choose the simplest and clearest format possible.

Make It Whole

A chart (or visualization of data) is rarely presented on its own. A dashboard, for example, may contain written content and several other charts. Always consider the context of your chart and design it to work within the larger whole.

Make It Sing

Trust your eye. There’s a form for your data that will lift it off the page and create the “Aha” moment for your audience. Sometimes you have to tweak your design until you find the right solution. The big picture and the details all matter.

Process

Visualizing Data is Hard

First, it’s important to acknowledge that this is not an easy thing to do and your first attempt to visualize your data may not be the best ultimate solution. Prepare to invest some time in trying a few different options before you build a final project. It will be worth the upfront time.

Ask Yourself, What Are the Goals of the Project?

Before you begin to focus too closely on the data, it’s important to be clear about the intent of the project. What is the purpose of presenting the data and how will your visualization contribute to the overall goals for the project?

Who Will Be Viewing or Using the Data?

It’s also important to know your audience. What do they care about in the overall context of your project and how are they most likely to connect to this data? What is the most critical information they need to see?

Is the Visualization Explanatory or Exploratory? Static or Interactive?

Most data visualization projects fall into these two categories: either presenting data as evidence of something being explained (typically in static form such as a PowerPoint presentation or printed report), or offering a presentation of data to allow a viewer to explore on their own (typically online in interactive form). An explanatory data visualization will require you to tell the viewer a story with the data, leveraging well-written chart headers and overlay notes to guide the viewer to the most important features of the data. An exploratory data visualization depends on your ability to provide the right kinds of interactive features that allow the view to discover the key features of the data on their own. 

Usability

Data visualizations have to present all informationclearly and concisely, prioritize information for understanding themat-a-glance and remove any possibility of ambiguities.Here are 5 simple tips that will help you increasethe usability of your charts:

Use Descriptive Captions

The caption of a chart should explicitly describe what the chart is all about. Consider creating a caption that is so descriptive that the axis labels (if used) are only left as reinforcements.

Arrange the Data in Order

If the data is name or category-based, and not time-based, it is a good idea to arrange the data in descending/ascending order.

This is particularly helpful when you have a lot of data sets to display on the chart. With arranged data, it is very easy to find out the best and worst in each category.

Delegate Detailed Info to Tooltips

Putting all the contextual data in a chart can make it overwhelming and hard to follow. Rather, consider ways to place the detailed information in interactive tooltips that appear on hover. This way, the user can choose to interact with the data to expose more information.

Add Context to Data

Data analysis becomes more meaningful when the data being plotted on the chart is given a context. This can be accomplished using trendlines. For example, showing a bar chart with a trendline for a sales goal helps the user immediately see whether the goal was met
each month.

Without the trendline, the chart would have just been showing the monthly sales for the year and you wouldn’t understand the context for the data.

The Z-Axis is Essential

Data is not two-dimensional and we should provide users with the opportunity to drill into the Z-axis. If a chart can be filtered, sorted or drilled down, provide the controls on the top or to the right and provide explicit instruction. When possible, provide the table view in your digital application in addition to the chart. This is useful for experts primarily interested in specific data values, as well as a dependable presentation of data for the visually impaired.