I'm surprised the techcrunch article got more points than the original writeup at http://waxy.org/2008/10/memeorandum_colors

I'm also pretty dissapointed this gets credited to me when it was mostly Andy's idea (I just did some math.)

Brendan O'Connor did a correlation between our output and a website that uses humans to measure skew at http://anyall.org/blog/2008/10/it-is-accurate-to-determine-a... which is pretty great. As a statistics and collective intelligence behavior junkie, his blog is pretty good reading, btw.

Andy's got a very popular blog, but your site is extremely well-known, and TechCrunch is a journalist site above all trying to get page hits.

Thanks for the link to Brendan's blog: I've never come across a social sciences blog that's this fun to read.

I don't suppose this is real-time, with continuously updated results? It seems you run once the SVD algorithm, store results in a spreadsheet, and then greasemonkey pulls the data from this static file.

Nope, I push them to Andy and he hand-checks them first. Could automate it, but maybe some other factor would be different someday?