Stat 5810, Applied Spatial Statistics
Homework Assignment 7 (11/14/00)
15 Points - Due Wed 11/29/00 in class or as a Web page
You may work in groups of 2 people on this assignment
and turn in a joint solution.
- 1) Start ArcView and work with the "Places" data set from
within ArcView. Select the "Variogram Cloud Link" and work
with a cutoff distance of 200000. Based on the variogram
cloud plots in XGobi, which of the 9 variables of the places data
show some spatial dependence? And which variables do not
show any spatial dependence? Use XGobi's smooth option
to support your answer.
Explore Climate, Arts, and Economy more carefully.
Are there any global outliers or any spatial outliers
in any of these three variables? It might be helpful to
run an independent XGobi with the Places data as well
to answer this question.
Incorporate meaningful screenshots (using xv). You should
use at most 10 different screenshots for this entire question.
Keep in mind that I print your answers in black and white.
So try to use significantly different colors and symbols
that are still distinguishable in a grayscale printout.
(5 Points)
- 2) Start ArcView and work with the "Satellite Image" data set from
within ArcView. Select the "Basic Link".
First remove the one outlier, then work with the grand tour
in XGobi (only include the 7 bands - do not include the ground
truth Sstruth2). Use the optimization based on "Holes".
Which crops are distinguishable from each other based on the
meaningful projections in the grand tour?
Carefully look at alfalfa. Is there anything specific about it?
And what about roads?
Again, support your answer by at most 5 meaningful screenshots.
(5 Points)
- 3) Revisit the Climate data from the "Places" data set within
S-Plus. The pure data is stored in
/home/symanzik/axx/xgobi/data_xgobi/places.
The corresponding column names are stored in
/home/symanzik/axx/xgobi/data_xgobi/places.col.
Read this data into S-Plus, construct a variogram based
on the method-of-moments and a robust variogram.
What are the range, sill, and nugget effect (if any)?
Then try to fit a theoretical variogram model.
(5 Points)