Spatial Statistics (ST 790 M), Fall 2006

PROJECT:

  • When you submit your project (if it is applied), give Dr. Fuentes a floppy with your data (so she can redo the analysis), in the floppy you should have only your RAW DATA, not the analysis. The project should be less than 10 pages length, and typed.
    DO NOT SEND DATA BY E-MAIL.
    The project can be based on a theoretical topic,  applied (based on an spatial analysis of real data) or mainly computational. Bayesian techniques are expected to be used.

If the project is applied make a thorough but concise report of your entire investigation, include at least the following:

  • Summary (goals and major findings)
  • Table of Contents
  • Description of the Reason for Your Study
  • Statement of How You A Priori Expected the Study to Turn Out
  • Explain the Raw Data
  • Appropriate Statistical Analyses of the Data (use graphs as well as numerical summaries): remove trend, exploratory analysis of sample semivariograms (lack of stationary??, anisotropy??), fit a theoretical semivariogram (WNLS, REML), maybe do kriging if prediction is the goal..
  • Statement of the Subject Matter Implications of Your Study, and
  • Discussion of Further Questions Raised by Your Study (that might be investigated in the future).

If you have a theoretical topic, you should follow similar steps, summary, table of contents, description of reason for your study, discussion...

If you have a computational topic, write a new SAS code (give the code in a floppy to Dr. Fuentes), to solve an interest problem in spatial statistics, you should applied your code to a simple dataset and include a discussion.

How to get interesting data for your project:
OZONE DATA (TOMS):
http://jwocky.gsfc.nasa.gov/
This is EPA's site for air quality data:
http://www.epa.gov/airprogm/airs/graphics/index.html
this is the site for acid rain data:
http://nadp.sws.uiuc.edu/nadpdata/
this site has some satellite data for NDVI, which measures plant characteristics. its a bit more complex, but should be interesting:
http://edcwww.cr.usgs.gov/glis/hyper/guide/usavhrr
and this is a place to start for weather data:
http://www.ncdc.noaa.gov/

HNO3 concentrations in ppm (parts per million) for July 11 1995, the columns of interest for you are the 4 (with col. number) the 5 (with row number) and the 7 (with HNO3 concentrations in ppm parts per million). Column 3 is the time of the day, you just could do for a fixed time, for a fixed value of column 3 (0 means 0 hours Greenwich central time). This is a very important pollutant, you could do an statistical analysis, check nonstationarity, anisotropy, study the spatial structure in small windows in your domain. You could also take the mean of the HNO3 values over time and study the spatial structure for this mean. COORDINATES: this file has 4 columns, column 3 is latitute in degrees, column 4 is longitude in degrees, column 1 is col. number and column 2 is row number for the HNO3 (nitric acid ) data. The data are in a regular grid with 75 columns and 69 rows (easter part of US). map (HNO3 july 11, 1995 at 17:00 GCT)

Back to st790 Home Page