I'm sitting on my own dock (nobody else here) on the mouth of the Cuyahoga River and Lake Erie.

I'm drinking my own beer unmolested even though this is 'prime entertainment real estate' in Cleveland's Flats on a gorgeous 82 degree and mostly sunny early evening.

A seagull just told me that "You're onto yourself."  But wtf do seagulls know?

If they were smart, seagulls wouldn't be hanging about a lake.

Except that the Great Lakes were initially mistaken as 'Sweet Seas' by early explorers due to their immensity and unpredictable volatility (and because they were 'seas' without salt).

So maybe the seagulls got taken in by the explorer hype.

No, no seagull pictures.  Seen one, seen them all.  (The pictures, not the seagulls themselves.)

Here's my setup.  Where are the pretty girls?  Where are all the drunken sailors?

Perhaps I should just let bygones be foregones and drink until I usher myself into a state of utter ecstasy.

Where goeth Xanga?

vertical axis: worldwide web-ranking by amount of traffic
                   (higher rank [lower number], say #1,
                    is better than lower rank [higher number], eg. #100)
horizontal axis: tick marks are 7 days apart

Measuring something repeatedly over time, such as the daily traffic ranking above, produces a time series.    The major task of time series analysis is to describe the nature of the past variation so that the future can be predicted and acted upon accordingly.

Although I didn't have the raw data, I extracted data points from the chart above and performed a rudimentary time series analysis.

Looking at the 3-month chart above, one can pretty clearly see a general trend upward toward higher ranking.  There is also an up-and-down pattern occuring quite regularly in the very short term.

Statistically, there is very strong evidence of trend with a strong 'seasonal' component.  The 'seasonal' component is the very short-term up-and-down pattern which in this case is a weekly fluctation with peaks on or about Saturday and valleys on or about Wednesdays.  Yep, Saturday's quake and humpday is Xanga's slump day!

Xanga's trend upward, based on the last 3 months, proves to be strongly linear.  That being so, I was able to devise a regression equation to describe it: y = -2.7(x) +126 .

(I've always wanted to reduce Xanga to manageable linearity!)

What this means is that, at present (based on the last 3 months  only), Xanga's rank improves (higher rank, lower number) by about -2.7 per week.  So if it were ranked 100th worldwide in terms of traffic today (which it is), then in one week is should be around 97th, and in a month around 89th.

There's my prediction.  But remember: I'm not a prophet.