It's that time again: CQ has posted the newest version of its yearly vote studies, ranking legislators on party unity and presidential support. Again, this uses my Flash applets for presenting the tabular data, as well as a scatter/distribution graphing.
As far as interesting emergent storylines go, there's not a lot for me to say yet. From the visualization end, I added medians to a couple of the plots but otherwise did relatively little tweaking. The one notable change was an adjustment to the House unity algorithm, due to the score of Rep. Walt Minnick, D-Idaho (and to a lesser extent, Rep. Bobby Bright, D-Alabama). Minnick has a unity score of 40%, the lowest of the House Democrats. As a result, I had to widen the "window" for that graph, which previously had no member with a unity score less than 50%. This had already been done in the Senate, thanks to Sen. Olympia Snowe, R-Maine.
You may notice some artifacting in the graph so far, particularly on the Democratic presidential support distribution. According to the editors for this data, it's probably due to the low amount of votes tallied for 2009 so far, causing a "clumping" around a few support values. As we accumulate more data and update these numbers, a more natural distribution curve should emerge.
My remaining technical gripes with these graphs, which I haven't had time to correct, are the confusing method of listing members in distribution views and the odd scaling that's used to fit them all in. I suspect they can both be solved by reducing the pixel size in those modes far enough that a 1:1 ratio is reached--no overlapping of values within columns. And I think we're going to take it widescreen, to make that easier--realistically, the whole thing's due for a design overhaul anyway. But in the meantime, I think it continues to work reasonably well, and it's still one of my favorite projects here.