Delight Don’t Distract: 3 Tips for Better Data Visualizations
Advocates for open data have long held that, although publishing government data is a critical first step in driving more transparency and accountability, simply making data available online is not an end in and of itself. Data must be curated and presented in easily accessible formats that everyday citizens can interpret and apply — not buried in dense tables or spreadsheets.
In fact, pioneers in the field have reached consensus that one of the most important tenets of open data is visualization. Stephen Goldsmith, Director of the Innovations in American Government Program at Harvard’s Kennedy School of Government, and a noted thought leader on this topic, recently asserted that: “…data without visualization is hardly open data at all.”
We know that data visualization is increasingly important. But what features separate the most effective charts, graphs, maps, and other displays from less useful or compelling ones? And how can you make sure that your data graphics lead to better comprehension, and not confusion? Here’s a quick set of tips to keep you and your team on track.
Recognize That Not All Data Visualizations Are the Same
Not all charts and graphs are equally suited for displaying the same types of data. Fortunately, there are some best practice guidelines for selecting the right visualization.
Seattle tracks the number of registered gardeners in community gardens
Bar charts work best when you have numerical data split out across multiple categories, and you want to compare quantities across those categories, such as budget allocations by department.
Riverside County uses data on internet users to build the case for county-wide broadband access
Pie charts are better suited to showing relative proportion, instead of direct comparisons.
St, Petersburg maps all requests reported through 311
Maps are, not surprisingly, the preferred option to plot geocoded data, such as incidents of crime by zip code.
There are numerous ways to visualize your data. Choosing the right format starts with evaluating the type of data you have and considering what you want the audience to take away from the data visualizations you present.
Mind the Labels
It may seem counterintuitive to a discussion on data graphics to talk about the significance of text. But words in the form of chart titles, descriptors for X and Y axes, data labels, legends and more are crucial to making your charts and graphs intelligible.
Let’s start with just a few of the basics.
Titles – Every data visualization needs a title. The title encapsulates the purpose of the graphic, gives clues to the information contained within, and signals the value it can deliver to the reader.
Axis Labels – For many types of graphs and charts, axis labels are essential to identify the variables being compared. Keep these labels as short as possible to avoid clutter, but descriptive enough to provide clarity at a glance.
Legends – Legends can be a particularly useful way to orient readers to charts that contain numerous color-code data categories, such as a scatter plot.
Although important to facilitating understanding, in all cases text should be as clear and direct as possible. Over-labeling (or what Edward Tufte refers to as “chartjunk”) can be distracting and lessen the impact of your visualization.
Go Beyond Visuals
Telling your story about your data is not limited to visualizations. There are numerous ways to contextualize data. For example, metadata, or information about the data itself, helps builds upon the concise bits of text used in chart titles, labels, and legends. It typically conveys information about the purpose of a specific data collection effort, how and when datasets were compiled, and limitations in scope. But analysts should strive to go even further, using visualizations to tell powerful, memorable stories.
Combining visualizations with narrative flow invites readers to engage and follow along on a deeper level, ask critical questions, and extract insights that prompt further discussion and discovery. Beyond these benefits, using visualizations to create colorful, multi-dimensional stories plays a role in increasing retention of the facts contained in the datasets. Readers are more likely to remember what they learned when your visualizations are woven into a thoughtful story.
See this example from the Washington Water and Salmon Fund Finder created by the Washington State Department of Ecology.
If you would like to gain more skills and rethink how you use data day-to-day, Socrata Education offers live and OnDemand courses on topics like data integration and site administration. Check out our course catalog.