Zack Abrahamson, the "data whiz" at political analysis site Politico, is apparently an R user. Politico's Feb 10 2012 chart of the day clearly uses the ggplot2 graphics package and (quoting Politico)
looks into the disenchanted slice of the GOP that’s not engaged with its party’s primary. And that slice doesn’t like Mitt Romney. People say turnout's down. When it's down, we actually see Romney stronger. [Minnesota is an exception.] This is a general election problem because the people who don't show up today won't show up in November. It’s the CPAC base. Here’s a chart comparing the change in turnout from 2008 to 2012 against the change in Romney support.
Those counties that had higher turnout in the 2012 primaries than in 2008 were associated with a comparatively lower support for Republican presidential candidate Mitt Romney. If I were to nitpick, the legend and color coding is a bit redundant here but, still: it's nice to techniques like small multiples, scatterplots and trensdlines used for political analysis.
(Note: the content of the link below changes daily. I couldn't find a permanent link for the Feb 10 Morning Report where the chart and quote above appeared.)
Politico: Morning Score (via @alexlundry)
How is Minnesota an exception? The scatterplot looks almost the same as Colorado. The regression lines are clearly weak, due to the amount of noise in the scatter plot. There is very little bivariate relationship between the variables. I think the results would be better conveyed as a set of boxplots for Change in turnout and another set of boxplots for Change in Romney support. My $.02.
Posted by: Kevin | February 10, 2012 at 09:24