I’m really excited to announce that I have just published an article about data visualization in the Winter 2014 edition of the Journal of Economic Perspectives (JEP). The JEP is one of the premier economics journals and I am thankful to the editors for inviting me to write the piece and for recognizing the growing importance of data visualization in the field. The following post is adapted—with some edits for space and context—from the published article.
When I think about the different types of data visualizations and the audiences they try to target, I think in spectrums or axes. On one axis, consider the connection between the two general forms of visualization.
- Static visualizations provide all of the information at once and are not active or moving: for example, visualizations that appear on paper are static.
- Interactive visualizations allow a transfer of information between the figure and the user. In general, the best visualizations in this category follow the three-step mantra laid out by Shneiderman (1996): “Overview first, zoom and filter, then details-on-demand.” Such visualizations give a broad look at the graphic space, then allow readers to further define the space of interest, and finally permit them to capture specific details.
Sitting somewhere between the two are animated visualizations. These types of visualizations—such as movies or online slideshows—do not necessarily permit the user to manipulate data points to create alternative results, but they do enable the user to control the pace of the story.
On the other axis, consider the function of the visualization.
- Explanatory visualizations bring the main results to the forefront—they “surface key findings”—to some extent helping to reveal the story (for further discussion, see Segel and Heer 2010; Kosara and Mackinlay 2013).
- Exploratory visualizations help users interact with a dataset or subject matter to uncover the findings themselves. Such visualizations do not generally propose a single narrative or draw out specific insights.
The intersection of the axes results in four quadrants:
- Static & explanatory: Graphs tend to be static and are used to reinforce a point made in accompanying text. The first thing many people probably think about in this space are standard charts: bars, lines, columns. Infographics often—but not always—reside in this space. Many people in my field of economics typically live in the world of explanatory, static graphs—they tend to be inserted in a report or article and used to reinforce a point made in accompanying text.
- Static & exploratory: These types of visualizations lead readers to discover their own stories as they examine the static representation of data. Moritz Stefaner’s Müsli Ingredient Network is a great example.
- Interactive & explanatory: Perhaps the easiest explanatory-interactive graph type to consider is a static graph that has an interactive hover or rollover layered on top (see, for example, these line graphs from the World Bank).
- Interactive & exploratory: These visualizations graphically present a complete data set and ask users to find interesting stories (for example, the OECD’s Better Life Index).
Other types of relationship mappings are obviously possible and examples include Kirk (2013), Kosara (2013), Heer, Bostock, and Ogievetsky (2010); Bertin (1983); Harris (1996).
You can use the relationship map I’ve described here to help you think carefully about how your audience will use your visualization and how it fits their needs.
