The page counter just turned 200.000 today, so I thought it was time to look at the site statistics again.
I've yet to fully understand the traffic burst aftter Xmas, since Technorati was apparently down (again), and I couldn't fathom if the Referrers I've been getting are due to more people posting random links or if the site is being actively linked from others.
Nevertheless, the trend of generating Spam Referrals continues, and after sending a couple more complaints to ISPs I've started testing a new approach to replace my quick and dirty hack. It's Bayesian, of course, and matches known "spam words" in the referrer URL, scoring it appropriately. It's also slow as hell at this point, so it's not really usable.
Another surprise came in when building a test file for it from the Referrers data: this link cropped up. Wow. Of course, only a few portuguese weblogs are ranked here, but I'm either first or second depending on average ranking, which is pretty sobering.
But then, most of my traffic is clearly not from Portuguese readers, so those numbers are essentially meaningless (a quick peek at the current forerunner revelaled two or three sites sharing the same counter, which further skews their meaning).
Doing The Math
In the meantime, more maths-oriented folk are sure to enjoy the little calculator I found today. The only thing it's lacking is graphing, but that's sure to be addressed eventually:
Clearing Out The Trash
I've also been clearing out old build trees from my archive server with this: