In the same spirit of exploration as for the Mike Cheponis fire,
examine the GCP data leading up to the shutdown of Fred Bosick's
egg due to a fire an a neighboring apartment.
Fred sent a note to let me know what had happened, and
suggested it might also be worth a look, and to add this
material to the exploratory database.
Date: Fri, 23 Jan 2004 15:11:36 [-0600 (CST)]
From: Fred
To: rdnelson
Subject: Egg 2222 is Down for an Unknown Time
Hi!
There was a fire in a neighboring apartment, and the
building is deemed unhinhabitable until inspection and
repair. My stuff, cat, computer, and REG is OK! So am
I. The computer lost power when the fusebox shorted
from soot. That event woke me up to find smoke in the
air. In other words, the fire way already out of
control when the REG stopped uploading bits! You might
want to examine my REG output. :-) I'm hoping to be
allowed to move back in and power up everything in a
week or so.
Thanks, Fred.
The last data packet from Egg 2222 arived at 20:05 on the 22nd.
For the exploratory analysis, I took the whole day up to 20:05, plotting
the cumulative deviation of scores for all eggs (red) for egg 2222
alone (green), and the difference. The first figure shows this,
and is actually a striking picture. The data from all eggs has a couple
of periods of persistent large deviation, including a strong negative
trend beginning about an hour before Fred's egg lost power. The second
figure shows this period separately, and displays more clearly the
extreme departure.
It is tempting to interpret this picture as showing a response to the
emergency situation, but when we look at egg 2222 alone, there is no
obvious trend -- the data show a classic random walk.
As we so often must observe, there is much to learn before we can draw
conclusions from the EGG data.
|