Visualizing time–The efficacy of lines

My intention in this blog entry is twofold: 1) to introduce a useful data visualization newsletter, and 2) to challenge an opinion that was stated in the current edition of the aforementioned newsletter. For some time now I’ve enjoyed the monthly edition of a newsletter written by Juan C. Dürsteler, published through his Web site at Mr. Dürsteler writes about topics that range across the entire data visualization spectum. His newsletter has often introduced me to organizations, sites, studies, and products that I hadn’t previously known. I appreciate his dedication to this work.

In the April issue of the newsletter, Mr. Dürsteler made a statement with which I disagree:

Representing the passing of time has its difficulties. Depicting space is relatively easy, since it can be bound to the concept of axis. An axis can be run in both directions and can appropriately reference the spatial dimension.

Time, nevertheless is not well represented by a line since it isn’t a spatial dimension and, moreover, it’s irreversible. Paradoxically, the widest used metaphor is that of the line. Maybe the key to it is considering time as just a state of space.[Emphasis mine.]

I believe that the use of a line to encode quantitative data in two-dimensional graphs with X and Y axes works extremely well for helping us to visualize, concpetualize, and make sense of values as they change through time. We naturally think of time as linear, thus using a line to depict it works well. Doing so in a horizontal manner from left to right (for those whose languages are written from left to right) is a convention we easily understand. Depicting increases in value as inclined slopes and decreases as downward slopes uses differences in 2-D position relative to a quantitative scale along the Y (vertical) axis in a way that even children can learn to interpret with little difficulty.

Certainly, no visual representation can depict time in all its complexity, but that’s hardly the point. When we use graphs to visualize time-series data, we’re doing something conceptually simple using a visual method that communicates just as simply. A line graph is an elegant solution that works incredibly well. There are occasions when other methods might be needed to depict time, but for making sense of and communicating most quantitative data, these occasions are infrequently required.

Despite my disagreement with Mr. Dürsteler on this particular point, I highly recommend his newsletter as a thought-provoking resource.


Comments are closed.