parseerror.com / vdqi
bare information graphics
vdqi is a graphing library. its goal is to lay data bare.
"Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away."-- Antoine de Saint-Exuper
gallery comparing original and vdqi-interpreted visualizations
a typical information graphic uses a tiny fraction of its ink to display the actual data; often the majority of is wasted on chartjunk.
the following graphics visualize the same data[1]:
which is most effective?
the first graphic may look more familiar to you, similar to those commonly seen in reports and presentations. step back and look at the graph as a whole — what we really see isn't the data; we see the style of the graph; gratuitious, heavy gridlines and 3D shadow. the grid lines convey little information; the bars' 3D effect convey absolutely none. when we begin to really think about it, the bars themselves are not strictly necessary.
vdqi contends the second is more effective, eliminating the chartjunk and displaying only the data. the relationship between the data is immediately clear; indeed, it is impossible to miss because there is very little else. closer inspection reveals that the points are rendered as their value, providing more specific information than the barchart, and using only a fraction of the ink.