What's your favorite data set?

I came here from the mass Reddit exodus and figured this would be as good a question as any.

I like Fisher's iris data set. It lends itself to a lot of classification techniques and since it's used pretty much everywhere, a statistical audience will generally be familiar with it.