01:10 pm, 24 Dec 04

### friendrank - full distribution

Here's LJ's top 25 (but again, keep in mind the bias in my data set). Right after I made that last post I realized who'd be #1, and it's not

And here's a graph of the distribution. For those of you who haven't peeked yet, user #1 is the lonely dot up there in the corner.

**jwz**. What's your guess?1 avva 689.312561 2 status 322.193115 3 kompressorpower 290.737885 4 jwz 237.453430 5 hypnox 232.653580 6 imomus 194.737915 7 kore 186.459412 8 ana 180.692673 9 spiritrover 178.347702 10 opportunitygrrl 164.877151 11 zoe_trope 162.871094 12 sinnocence 159.299973 13 masuimi 147.922989 14 gnomeygirl 147.293320 15 mylastsigh 140.772308 16 edbook 136.235352 17 brad 135.409088 18 stereolabrat 127.292786 19 motel666 127.073608 20 jacqui 126.502151 21 billiam 125.090950 22 fumanchaw 124.384567 23 touchyou 122.320450 24 theashifaction 120.748344 25 spoonfeeding 107.679520

And here's a graph of the distribution. For those of you who haven't peeked yet, user #1 is the lonely dot up there in the corner.

xaosenkosmos(and i'm still trying very hard to make my poor little head simultaneously linearize the axes of that graph. it hurts.)

evan% (for i in *; do echo -n $i; grep '< ' $i | wc -l; done) > friendofs% sort -k 2 -n -r friendofs | head -25

status 2445

avva 2308

jwz 1476

kompressorpower 1124

spiritrover 1098

hypnox 1090

opportunitygrrl 1053

kore 1013

ana 960

zoe_trope 931

brad 832

gnomeygirl 821

mylastsigh 795

imomus 787

jacqui 701

fumanchaw 679

sinnocence 672

motel666 664

pjammer 644

satirical 622

masuimi 614

edbook 608

billiam 605

spoonfeeding 596

stereolabrat 592

pagerank has a component which gives effectively gives every edge some minimum score; the intuition is that it models random jumps by websurfers.

on my to-do list is to make that parameter tweakable via the command line. it's mostly implemented; i'll report back with more info.

when log-log graphs are a straight line, then the linear-linear looks like 1/x.

evanpbrane## slight nitpick:

when log-log graphs are a straight line, then the linear-linear looks like 1/x.since *this* log-log graph is a straight line, then it's linear-linear plot is a 1/x (1/sqrt(x) I'd guess, actually) - in general it's a monomial with degree equal to the slope (this one has slope roughly -1/2 or so.

cool stuff you got here, very neat!

evanscottbateman