25 October 2007
Since I saw Fernanda Viegas and Martin Wattenberg's presentation on Many Eyes a few weeks ago in our Applied Stats workshop, I've been itching to use their visualization tools on some of my own data. Tonight I made a treemap of the dataset of UK politicians that Jens Hainmueller and I have been developing. (The data consists of over 6000 candidates who ran for the House of Commons between 1950 and 1970.) I set up the visualization such that each box in the treemap is sized to indicate the number of campaigns for each combination of party and occupation (eg Conservative barristers) and the color reflects "proportion attending Oxbridge." But you can play with it via the menus at the bottom of the visualization and cut the data the way you want: you can make the size reflect "proportion female" and the color reflect "proportion elected," and you can make it show party by occupation instead of occupation by party. I've embedded the visualization here; you can make comments tied to a particular view of the data on the many-eyes site. I'm eager to hear reactions, whether on the visualization or on the brand-new data.