Bommarito and Katz’s visualization of Senate campaign contributions is a wonderful demonstration of the potential for making large amounts of government information accessible.
Having focused successively on campaigns to improve literacy, then media literacy, then design literacy, compelling work like this suggests that the time may soon come for "visualization literacy": the skills a citizen needs to evaluate the information presented via data visualizations.
Maps are a long-established visualization tool. They may purport to be value neutral, but of course are not. Depending on who’s paying, some maps show churches, and others show hotels. Every map embodies decisions to show some things and hide others, which means that every map is trying to persuade us to see the world in a certain way.
In the same way, more recent data visualizations have a more or less hidden agenda by virtue of their choices of what to show, and how to show it.
Looking deeper, the data being visualized is itself a subset of all possible information, and its collection is based on a model of the world that entails various biases. Data has an agenda, not only in what was collected, but in how it was gathered, and how it is categorized.
Finally, the interface that one has to use to access the data ("interface" as the set of commands and messages, not the visuals) has an agenda. If software code implements an architecture, as Larry Lessig would have it, then the code interfaces place bounds on the architecture. Even if all data were accessible with a given interface, some things will always be easier to do than others. This biases not only the data that is represented, but also the ways in which it is represented.
Information at every level – from data selection, to interface calls, to visualization – comes with an agenda. By the “if I can imagine it, it’s already been done” rule, seminars in the semiotics of visualization have surely already sprung up.
Update 4 May 2009:
Courtesy SiliconValley.com, here's a recent example of how maps carry meaning: Historical maps of Japan from Berkeley's East Asian Library collection that were offered as layers that could applied in Google Earth have had to be sanitized to remove references to burakumin villages. David Rumsey, who oversees the Berkeley collection, acknowledged, "We tend to think of maps as factual, like a satellite picture, but maps are never neutral; they always have a certain point of view."