|
First reports of the tragedy in Myanmar indicated hundreds
dead and thousands homeless. Here is an extract from early
CNN news coverage:
'Residents in this sprawling river delta
city hacked their way through downed trees and trudged
through knee-deep swirling brown waters Monday as they tried
to pick up the pieces of their lives after a deadly cyclone
ravaged the southeast Asian country over the weekend.
The powerful storm hit hard on Saturday, May 3, but had been
building offshore over the previous day. The tropical cyclone,
packing winds of up to 150 miles (241 km) per hour, slammed
into Myanmar over the weekend, killing as many as 350 people.
"We believe hundreds of people are dead," said Khin Maung
Win with the Democratic Voice of Burma -- a broadcast media
group run by opposition expatriates. "The entire lower Burma
is affected. In some areas, entire villages are wiped out."'
Sadly, this early reporting has been supplanted with
estimates of 100,000 dead, and millions displaced and in
need of food, shelter, and above all, clean water. Disease
is now the primary concern, with cholera already taking
hold.
The GCP prediction was made based on the early reports, and
focused on May 3, using the full GCP day as the formal data
set. The result is a modest deviation, with Chisquare
86485.089 on 86400 df, yielding a p-value of 0.418 and corresponding
Z-score of 0.206.
The graphic display using the cumulative deviation of
Chisquare shows a strong trend from early in the morning
until about 1600, when it reverses to become an equally
strong negative trend for the rest of the day.
Because this event has continued and worsened, it seems
useful to look at the data for the ensuing days. The
following figure shows three days beginning with the 3rd of
May. The negative trend continues. This is of course a post
facto analysis (and, moreover, we know that the data are
generally too noisy for reliable interpretation of
individual event analyses) but one might say the picture is
symbolically appropriate to the situation in Myanmar.
|