ST590G -- FALL 2012 
     
  Class Exercise #4 -- Thursday, 22 September 2012

  Our Group ___________________________________________________

  The Norwegian snow data one more time.  In the directory "snow" 
  are files from many sites in Norway containing dates and snow depths 
  recorded daily for many years.  A subset of the files include:

     asker.txt  bjornholt.txt  kjobli.txt  listafyr.txt oslo.txt

     songli.txt  stavanger.txt tafjord.txt  tromso.txt  utsira.txt

  In one previous exercise (#2), you looked at some statistics for each 
  snow year and determined what made a 'good' year.  For some, the criteria
  were 'no more than 200 zero and no more than 50 missing.'  In the other 
  exercise (#3), you found the largest observation and the date  when it 
  occurred.  In the first case, you can construct a dataset that includes 
  snowyears and a variable designating it 'good' or 'not good.'  In the 
  second case, you have a dataset with snowyear, maxdepth, and maxdate.

  1) Recode your work from Exercise #2 to get snowyear, maxdepth, and 
  whatever you need to determine a 'good' or 'not good' snowyear.

  2) Regress the maxdepth on snowyear and create a dataset with a single 
  observation with the slope estimate (and its se etc.).  

  3) Read the lat, lon, and altitude information from the header of the
  file.  Merge this single observation with your information on the slope
  estimate.

  4) Write a macro to do all of this, taking the station dataset name as 
  its argument and producing a single observation SAS dataset.
  (We will then look at ways to automate this further.)

  Submit your output and log.  Don't forget your toes!