We have many dashboards available to us today in many contexts: finance, forecasts, consumption, demographics, etc. We...
We have many dashboards available to us today in many contexts: finance, forecasts, consumption, demographics, etc. We witness the emergence of beautiful graphical representations more frequently than ever. Instead of flooding this post with examples of data visualization, I would like to put the spotlight on how we processes them, and what challenges that presents.
A couple of years ago, we were tasked at IDEO to design dashboard visualizations for Ford's next generation hybrid vehicle, the Ford Fusion. Hybrid cars are efficient only if the driver maximizes the car's potential; in other words, if the driver learns to make sense of the complex mechanics of two motors and regenerative braking. If the dashboard is the interface between the driver and the car, how might it coach drivers to make sense of this complexity and to adopt efficient driving habits? That is the obvious question, but it's incomplete. The missing part is, how do we design without interfering with driving and safety?
This is the key question we need to address when we put people at the center of evaluation (as Jocelyn Wyatt puts it). In this context, it means recognizing that people are preoccupied with more important tasks than spending long amounts of time in front of dashboards and data visualizations. This is true in any setting, and in our case it was driving. The role of visualization should not be to demand full attention, but to support the priority task and improve it through feedback loops. The challenge is not just to display how you are doing right now, but also to figure out how you could do better. So, what does this mean for the visualization itself?
Every form of visualization should tell a story. Unfortunately there is limited attention and time to process all the stories. So the gist of the story, or its immediate impact, should be visible right away. The term I like to use for this principle is "glanceability." What does a visualization tell us before we take time to analyze it? I invite you to look at the following chart and image for 10 seconds each and compare. What did you see? What did you feel?
Modified from Azar Askin's reproduction of a poster by Muenster Planning Office, Germany
Displaying too many numbers upfront can be overwhelming, yet that does not mean we should avoid them altogether when visualizing data. Pushing the narrative metaphor further, the story of visualizations should unfold as one reads into them.
Al Gore highlights the dramatic increase in CO2 concentration by stepping on an elevated platform in the movie An Inconvenient Truth.
Today, we can use simple interactivity to accommodate an increased level of detail on demand. In more advanced terms, visualizations should allow for direct manipulation as Ben Shneiderman, one of the pioneers of information visualization field, put it in 1983.
Demographic data is often presented in very complex tables and boring charts. Gapminder presents historical data related to poverty in visually compelling and easily maneuverable ways, and actually confronts many of the myths we hold about the developed and developing worlds through the narrative told using the data.
You can explore more examples of data visualization in web and print publications such as Information Aesthetics, Flowing Data, Many Eyes, Wired, The New York Times, and right here at GOOD.
Some questions for further discussion:
- Do you remember a visualization that told you a story?
- What did you see at first glance?
- How did you dig deeper?
Engin Erdo?an designs interactions and leads projects at IDEO.
UPDATE: We've removed an image previously posted at the request of the author.