[:0] [b]HOLD EVERYTHING!!![/b] [:0]
Things just weren't looking right to me on the original correlation chart, so I went back and dug deeper. Apparently, having so many null values in the original comparison breaks the Spearman Rank Correlation algorithm...who knew?! [:o)] It's like trying to compare Apples and Oranges since everyone essentially was judged on different song sets. Or maybe like having some people pick amongst a bunch of vegetables and others amongst different meats of the world, then asking how they were related to eachother. I needed to add a real "0" to each of the null values instead - all fixed!
Here's the new chart that makes MUCH more sense. We're seeing almost everyone break off into two main clusters, we'll call them "L" and "R" - kinda nifty. Then there's those freaks Eight6, Opera Prima and YipYip way off on their own by a far margin, we'll call them cluster "X". There's absolutely zero correlation between them and anyone else! [:p]
Both L and R clusters seem to further segregate into several subclusters. A lot of things stayed the same, many things changed quite a lot. In those clusters, L4 and R1 are the most distinct from the rest of their respective clusters, each falling into their own root sub-cluster before the other clusters branch off themselves.
The most related individuals were Fanciz and Rubbeet. There must be something in the water...something BAD, LOL! TomD is by far the most related to the mean of the NWO represented here, with MikeP also in that group - interesting. I'm proud to be cozy with Cubby - go BSP bro!!! Manekineko now falls a little farther out but not by much. The Long Island inbreeding is apparent in Abs, Marv and BigBri's clustering - cool.
You guys see anything else? Happier with this model?? [8D]
<img src="http://www.pbase.com/myirwin/image/68751750/original.jpg">

