In linguistics, we often make use of charts as part of our analysis and argument. And this is true of any field that makes use of quantitative data. The importance of how that data is represented is, I think, an often under-examined piece of argumentation. Charts and graphs can be enormously persuasive, and we have powerful tools at our disposal for creating different kinds of data displays. Of course, like any rhetorical tool, context matters. Increasingly beautiful and increasingly complex is not always the better option. That said, a gray-scale bar chart isn’t always the best option either. An author needs to think through purpose and audience.
For some really interesting perspective of these issues, I found this tumblr, which is maintained by some of the creators of the New York Times infographics. It walks through both conceptual and design issues related to visual data representation. Very cool.
This entry is particularly useful, as it presents different iterations of the same chart and discusses their various strengths and weaknesses.
Also, the image at the top of this entry is a famous one created by Florence Nightingale, who was an accomplished statistician in addition to being a nurse. For more information on her chart, check here.