For this week’s ITConversations podcast I spoke with Fernanda Viégas and Martin Wattenberg about Many Eyes, a project at the forefront of a new category called social data visualization. I was particularly interested to hear about how civic or political argumentation, which tends to devolve into posturing — especially online — might improve when it’s grounded in shared data.
I don’t think we’ve reached data analysis utopia, but there are intriguing first steps. We’ve had a couple of solid political arguments happen on the site, where someone will put up a vehemently argumentative piece, saying for example that they believe people on welfare get too much money, and they’ll point to their statistics and charts to support that. In the skirmish that follows, people often get beyond the standard red-state/blue-state divide because it is rooted in the numbers.
Of course this argumentation doesn’t happen only on the Many Eyes site. Because the visualizations are linkable and (now) embeddable, it spills over to blogs as well.
I’m also intrigued by the notion that, as more people spend more time investigating official sources of data, we’ll start to uncover problems with the quality of that data.
We’ve seen that people have found mistakes on official data sets from authoritative sources. And the reason they were able to do that so easily is because visualization will show you something you didn’t have time or patience to discover in a spreadsheet.
This does leave Many Eyes open to the criticism that it invites people who lack statistical expertise to draw fallacious conclusions.
We’ve certainly run into objections that visualization can be deceptive. People are afraid that visualizations will be created that are inappropriate and misleading. And in fact that’s a well-founded objection in some ways, because every visualization is a simplification of the underlying data. There’s a point of view involved, and people are suspicious of that. But my belief is we have to give people as much power as we can.
Also, even though we created Many Eyes for the lay person, because we felt that this was something that was needed, that there was nothing out there for people to play with in terms of interactive visualization, it seems to be powerful enough to attract scientists too.
I think that kind of cultural mashup will be really good for everyone involved. Fernanda talked about how, at one point, a user put up a visualization of the Alberto Gonzales testimony that highlighted lots of “I don’t recall” kinds of statements. An hour later, another user put up a similar visualization of Bill Clinton’s testimony about Monica Lewinsky.
You could see the same sorts of phrases. The really exciting part, to us, was a couple of things. One, there was a conversation going on through these visualizations. Also, people usually think of informational visualization as this neutral tool, because it is based on data. Part of what people are beginning to understand, hopefully, is that visualizations have a point of view.